Efficient reference structures are essential in video compression, enabling the exploitation of temporal dependencies across frames to reduce redundancy. In this paper, we delve into the inter-frame reference management mechanism in neural video codecs (NVCs). Previous schemes have inherited the reference propagation mechanism with the guidance of predefined reference structure, but the reference modeling across diverse reference sources remains underexplored. Moreover, the mismatch between the reference structure used for motion estimation and motion compensation limits the effectiveness of inter-frame prediction. To address the above limitations, we propose the unified reference hierarchy that integrates a learned hierarchical reference structure into the existing inherent reference propagation mechanism. Specifically, we first propose the hierarchical reference structure (HRS) to manage the multiple temporal contexts in the propagated reference feature, where a hierarchy-aware reference modulation module is integrated to select the most relevant reference features across different quality levels under the guidance of the reference balance loss. In addition, we propose the HRS-guided feature-wise inter-frame prediction that learns the low-rank approximation of the selected reference feature for ensuring the consistency and improving the inter-frame prediction performance. We conduct experiments on a state-of-the-art NVC, DCVC-DC. Experimental results show that our codec achieves an average 26% bitrate saving over H.266/VVC, and a 28.2% bitrate reduction compared to DCVC-DC without increasing the decoding complexity.

Embedding deep neural networks (NNs) into mixed-integer programs (MIPs) is attractive for decision making with learned constraints, yet state-of-the-art monolithic linearisations blow up in size and quickly become intractable. In this paper, we introduce a novel dual-decomposition framework that relaxes the single coupling equality u=x with an augmented Lagrange multiplier and splits the problem into a vanilla MIP and a constrained NN block. Each part is tackled by the solver that suits it best-branch and cut for the MIP subproblem, first-order optimisation for the NN subproblem, so the model remains modular, the number of integer variables never grows with network depth, and the per-iteration cost scales only linearly with the NN size. On the public SurrogateLIB benchmark, our method proves scalable, modular, and adaptable: it runs 120x faster than an exact Big-M formulation on the largest test case; the NN sub-solver can be swapped from a log-barrier interior step to a projected-gradient routine with no code changes; and swapping the MLP for an LSTM backbone still completes the full optimisation in 47s without any bespoke adaptation.

The clustering-based placement framework has demonstrated promising potential in improving the efficiency and quality of very-large-scale integration (VLSI) placement. However, existing methods typically impose unified and rule-based constraints on different clusters, overlooking the unique intra-and inter-cluster connection properties that vary across clusters, which leads to suboptimal results. To address this challenge and promote effective PPA optimization, we introduce an innovative PPA-driven placement paradigm with mixed-grained Adaptive Cluster Constraints Optimization (ACCO), which applies constraints with customized constraint tightness to different clusters, balancing local and global interactions for improved placement performance. Specifically, we propose a novel eBound model with quantified constraint tightness, combined with a Bayesian optimizer to dynamically adjust the constraints for each cluster based on PPA outcomes, which are ultimately passed on to the final flat placement. Experimental results on benchmarks across various domains show that our methods can achieve up to 62%, 97% and 25% improvements in post-route WNS, TNS and power compared to existing methods.

Timely and personalized treatment decisions are essential across a wide range of healthcare settings where patient responses can vary significantly and evolve over time. Clinical data used to support these treatment decisions are often irregularly sampled, where missing data frequencies may implicitly convey information about the patient's condition. Existing Reinforcement Learning (RL) based clinical decision support systems often ignore the missing patterns and distort them with coarse discretization and simple imputation. They are also predominantly model-free and largely depend on retrospective data, which could lead to insufficient exploration and bias by historical behaviors. To address these limitations, we propose medDreamer, a novel model-based reinforcement learning framework for personalized treatment recommendation. medDreamer contains a world model with an Adaptive Feature Integration module that simulates latent patient states from irregular data and a two-phase policy trained on a hybrid of real and imagined trajectories. This enables learning optimal policies that go beyond the sub-optimality of historical clinical decisions, while remaining close to real clinical data. We evaluate medDreamer on both sepsis and mechanical ventilation treatment tasks using two large-scale Electronic Health Records (EHRs) datasets. Comprehensive evaluations show that medDreamer significantly outperforms model-free and model-based baselines in both clinical outcomes and off-policy metrics.

Adversarial examples are well known to pose a security risk, when attacking deep learning models. While, most of existing adversarial attacks are designed to attack a single deep learning-based task, such as image classification. In practical scenarios, it is more necessary to study adversarial examples transferring across different vision tasks. However, it is challenging to create cross-task adversarial examples that can destroy multiple vision tasks at once due to unavailable various task-specific models and loss functions for attackers. To deal with this problem, we propose a Dual Attention-Guided Method (DAGM) for crafting cross-task adversarial examples by designing a spatial attention module and a channel attention module to capture overlapping discriminative regions and features that contribute to various tasks. Then we craft cross-task adversarial examples via reducing the dispersion (i.e., standard deviation) of feature maps re-weighted by both attention modules, which can destroy the overlapping discriminative regions and features for various tasks. Furthermore, to present theoretical explanation, we systematically analyze our method, and rigorously prove that both attention modules can provide better effectiveness of our adversarial examples, compared with existing cross-task adversarial attacks. Extensive experiments on two datasets demonstrate that our method can significantly degrade the performance of various tasks, even online CV APIs, and consistently outperform state-of-the-art methods by a large margin.

In interactive point cloud segmentation, users can achieve higher accuracy object masks than in instance segmentation by performing limited positive and/or negative clicks on the objects of interest in the scene. Existing methods often employ sparse click representations, leading the model to focus more on local detail features around the click points and failing to fully exploit the guidance information provided by each click, thus impacting the click effectiveness. We utilize a dense representation that reflects spatial distance relationships, known as the distance map, as the click channel to tackle the sparsity problem of click representation in current approaches. Based on the distance map, we introduce ClickEnhance, which is designed to maximize the guiding impact of each click. The proposed method encompasses the design of a click-specific encoder and the utilization of contrastive learning. The Click-Specific Encoder ensures that the network can adequately consider the influence of individual clicks during the feature encoding phase. Contrastive learning, on the other hand, reduces the feature distance between the click points and the target object, thus simplifying the subsequent segmentation process. Experimental results demonstrate that the ClickEnhance method markedly improves segmentation performance across multiple datasets, exhibiting superior generalization capabilities on challenging datasets compared to the state-of-the-art methods. This allows for the generation of high-precision object-level masks with fewer interactions, indicating great potential for practical applications.

Column Generation (CG) is an effective and iterative algorithm to solve large-scale linear programs (LP). During each CG iteration, new columns are added to improve the solution of the LP. Typically, CG greedily selects one column with the most negative reduced cost, which can be improved by adding more columns at once. However, selecting all columns with negative reduced costs would lead to the addition of redundant columns that do not improve the objective value. Therefore, selecting the appropriate columns to add is still an open problem and previous machine-learning-based approaches for CG only add a constant quantity of columns per iteration due to the state-space explosion problem. To address this, we propose Fast Family Column Generation (FFCG) — a novel reinforcement-learning-based CG that selects a variable number of columns as needed in an iteration. Specifically, we formulate the column selection problem in CG as an MDP and design a reward metric that balances both the convergence speed and the number of redundant columns. In our experiments, FFCG converges faster on the common benchmarks and reduces the number of CG iterations by 77.1% for Cutting Stock Problem (CSP) and 84.8% for Vehicle Routing Problem with Time Windows (VRPTW), and a 71.4% reduction in computing time for CSP and 84.0% for VRPTW on average compared to several state-of-the-art baselines.

The key to semi-supervised semantic segmentation lies in how to fully exploit a large amount of unlabeled data to improve the model’s generalization performance. Most methods are lured into the trap of taking each class independently (i.e., class-independent consistency) and neglecting the fact that there exist semantic dependencies among classes. In this paper, we analyze the bottlenecks of class-independent consistency inherent in previous methods and offer a fresh perspective of cooperative game theory to explicitly encourage class-consensus alignment (i.e., class-consensus consistency between the teacher (weak augmented view) and student network (strong augmented view). We formulate classes as players in an cooperative game to model their interpretable consensus and shed light on the possibility of closer collaboration between consensus themselves and consistency regularization, yielding more comprehensive and effective supervision signals. To this end, we carefully design the class-consensus consistency without introducing any external knowledge to model class structure information which renders better interpretability, and further, prepend relaxed class-consensus consistency (RCC) to unlock the potential of modeling class consensus by relaxing the strict alignment of direct class consensus values to ranking alignment. Extensive experimental results on multiple benchmarks demonstrate that RCC performs favorably against state-of-the-art methods. Particularly in the low-data regimes, RCC achieves significant improvements.

Relative pose estimation provides a promising way for achieving object-agnostic pose estimation. Despite the success of existing 3D correspondence-based methods, the reliance on explicit feature matching suffers from small overlaps in visible regions and unreliable feature estimation for invisible regions. Inspired by humans’ ability to assemble two object parts that have small or no overlapping regions by considering object structure, we propose a novel Structure-Aware Correspondence Learning method for Relative Pose Estimation, which consists of two key modules. First, a structure-aware keypoint extraction module is designed to locate a set of kepoints that can represent the structure of objects with different shapes and appearance, under the guidance of a keypoint based image reconstruction loss. Second, a structure-aware correspondence estimation module is designed to model the intra-image and inter-image relationships between keypoints to extract structure-aware features for correspondence estimation. By jointly leveraging these two modules, the proposed method can naturally estimate 3D-3D correspondences for unseen objects without explicit feature matching for precise relative pose estimation. Experimental results on the CO3D, Objaverse and LineMOD datasets demonstrate that the proposed method significantly outperforms prior methods, i.e., with 5.7° reduction in mean angular error on the CO3D dataset.

The rapid advancement of large-scale deep neural networks and large language models has intensified the demand for highly efficient GPU clusters. However, existing distributed training frameworks, like Fat-tree and TopoOpt, struggle with inefficient resource utilization and network bottlenecks. They often optimize communication, parallelism, and network topology independently, failing to leverage their interdependencies. To address this gap, we propose TOPO-X, a novel reconfigurable network framework that co-optimizes flow scheduling, training parallelism, and optical network topology. By formulating this integrated optimization challenge as a Resource-Constrained Project Scheduling Problem, TOPO-X dynamically adapts to changing workloads and network conditions using optical network reconfiguration capabilities. Our experimental results show that TOPO-X outperforms the state-of-the-art solution, TopoOpt, achieving a 2.22× speedup in training iteration times on average. These findings highlight TOPO-X as a promising approach for scalable, adaptive, and high-performance GPU clusters designed to meet the increasing demands of large-scale AI training workloads.

Leveraging machine learning (ML) to predict an initial solution for mixed-integer linear programming (MILP) has gained considerable popularity in recent years. These methods predict a solution and fix a subset of variables to reduce the problem dimension. Then, they solve the reduced problem to obtain the final solutions. However, directly fixing variable values can lead to low-quality solutions or even infeasible reduced problems if the predicted solution is not accurate enough. To address this challenge, we propose an Alternating prediction-correction neural solving framework (Apollo-MILP) that can identify and select accurate and reliable predicted values to fix. In each iteration, Apollo-MILP conducts a prediction step for the unfixed variables, followed by a correction step to obtain an improved solution (called reference solution) through a trust-region search. By incorporating the predicted and reference solutions, we introduce a novel Uncertainty-based Error upper BOund (UEBO) to evaluate the uncertainty of the predicted values and fix those with high confidence. A notable feature of Apollo-MILP is the superior ability for problem reduction while preserving optimality, leading to high-quality final solutions. Experiments on commonly used benchmarks demonstrate that our proposed Apollo-MILP significantly outperforms other ML-based approaches in terms of solution quality, achieving over a 50% reduction in the solution gap.

The efficiency of Logic Optimization (LO) has become one of the key bottlenecks in chip design. To prompt efficient LO, previous studies propose using a key scoring function to predict and prune a large number of ineffective nodes of the LO heuristics. However, the existing scoring functions struggle to balance inference efficiency, interpretability, and generalization performance, which severely hinders their application to modern LO tools. To address this challenge, we propose a novel data-driven circuit symbolic learning framework, namely CMO, to learn lightweight, interpretable, and generalizable scoring functions. The major challenge of developing CMO is to discover symbolic functions that can well generalize to unseen circuits, i.e., the circuit symbolic generalization problem. Thus, the major technical contribution of CMO is the novel Graph Enhanced Symbolic Discovery framework, which distills dark knowledge from a well-designed Graph Neural Network (GNN) to enhance the generalization capability of the learned symbolic functions. To the best of our knowledge, CMO is *the first* graph-enhanced approach for discovering lightweight and interpretable symbolic functions that can well generalize to unseen circuits in LO. Experiments on three challenging circuit benchmarks show that the *interpretable* symbolic functions learned by CMO outperform previous state-of-the-art (SOTA) GPU-based and human-designed approaches in terms of *inference efficiency* and *generalization capability*. Moreover, we integrate CMO with the Mfs2 heuristic---one of the most time-consuming LO heuristics. The empirical results demonstrate that CMO significantly improves its efficiency while keeping comparable optimization performance when executed on a CPU-based machine, achieving up to 2.5× faster runtime.

Despite the impressive performance of general large language models(LLMs), many of their applications in specific domains (e.g., low-data and knowledge-intensive) still confront significant challenges. Supervised fine-tuning (SFT)---where a general LLM is further trained on a small labeled dataset to adapt for specific tasks or domains---has shown great power for developing domain-specific LLMs. However, existing SFT data primarily consist of Question and Answer (Q&A) pairs, which poses a significant challenge for LLMs to comprehend the correlation and logic of knowledge underlying the Q&A. To address this challenge, we propose a conceptually flexible and general framework to boost SFT, namely Knowledge Graph-Driven Supervised Fine-Tuning (KG-SFT). The key idea of KG-SFT is to generate high-quality explanations for each Q&A pair via a structured knowledge graph to enhance the knowledge comprehension and manipulation of LLMs. Specifically, KG-SFT consists of three components: Extractor, Generator, and Detector. For a given Q&A pair, (i) Extractor first identifies entities within Q&A pairs and extracts relevant reasoning subgraphs from external KGs, (ii) Generator then produces corresponding fluent explanations utilizing these reasoning subgraphs, and (iii) finally, Detector performs sentence-level knowledge conflicts detection on these explanations to guarantee the reliability. KG-SFT focuses on generating high-quality explanations to improve the quality of Q&A pair, which reveals a promising direction for supplementing existing data augmentation methods. Extensive experiments on fifteen different domains and six different languages demonstrate the effectiveness of KG-SFT, leading to an accuracy improvement of up to 18% and an average of 8.7% in low-data scenarios.

Machine learning techniques have shown great potential in enhancing macro placement, a critical stage in modern chip design. However, existing methods primarily focus on *online* optimization of *intermediate surrogate metrics* that are available at the current placement stage, rather than directly targeting the *cross-stage metrics*---such as the timing performance---that measure the final chip quality. This is mainly because of the high computational costs associated with performing post-placement stages for evaluating such metrics, making the *online* optimization impractical. Consequently, these optimizations struggle to align with actual performance improvements and can even lead to severe manufacturing issues. To bridge this gap, we propose **LaMPlace**, which **L**earns **a** **M**ask for optimizing cross-stage metrics in macro placement. Specifically, LaMPlace trains a predictor on *offline* data to estimate these *cross-stage metrics* and then leverages the predictor to quickly generate a mask, i.e., a pixel-level feature map that quantifies the impact of placing a macro in each chip grid location on the design metrics. This mask essentially acts as a fast evaluator, enabling placement decisions based on *cross-stage metrics* rather than *intermediate surrogate metrics*. Experiments on commonly used benchmarks demonstrate that LaMPlace significantly improves the chip quality across several key design metrics, achieving an average improvement of 9.6\%, notably 43.0\% and 30.4\% in terms of WNS and TNS, respectively, which are two crucial cross-stage metrics that reflect the final chip quality in terms of the timing performance.

Machine learning (ML) techniques have shown great potential in generating high-quality solutions for integer linear programs (ILPs). However, existing methods typically rely on a *supervised learning* paradigm, leading to (1) *expensive training cost* due to repeated invocations of traditional solvers to generate training labels, and (2) *plausible yet infeasible solutions* due to the misalignment between the training objective (minimizing prediction loss) and the inference objective (generating high-quality solutions). To tackle this challenge, we propose **DiffILO** (**Diff**erentiable **I**nteger **L**inear Programming **O**ptimization), an *unsupervised learning paradigm for learning to solve ILPs*. Specifically, through a novel probabilistic modeling, DiffILO reformulates ILPs---discrete and constrained optimization problems---into continuous, differentiable (almost everywhere), and unconstrained optimization problems. This reformulation enables DiffILO to simultaneously solve ILPs and train the model via straightforward gradient descent, providing two major advantages. First, it significantly reduces the training cost, as the training process does not need the aid of traditional solvers at all. Second, it facilitates the generation of feasible and high-quality solutions, as the model *learns to solve ILPs* in an end-to-end manner, thus aligning the training and inference objectives. Experiments on commonly used ILP datasets demonstrate that DiffILO not only achieves an average training speedup of $13.2$ times compared to supervised methods, but also outperforms them by generating heuristic solutions with significantly higher feasibility ratios and much better solution qualities.

We introduce D-FINE, a powerful real-time object detector that achieves outstanding localization precision by redefining the bounding box regression task in DETR models. D-FINE comprises two key components: Fine-grained Distribution Refinement (FDR) and Global Optimal Localization Self-Distillation (GO-LSD). FDR transforms the regression process from predicting fixed coordinates to iteratively refining probability distributions, providing a fine-grained intermediate representation that significantly enhances localization accuracy. GO-LSD is a bidirectional optimization strategy that transfers localization knowledge from refined distributions to shallower layers through self-distillation, while also simplifying the residual prediction tasks for deeper layers. Additionally, D-FINE incorporates lightweight optimizations in computationally intensive modules and operations, achieving a better balance between speed and accuracy. Specifically, D-FINE-L / X achieves 54.0% / 55.8% AP on the COCO dataset at 124 / 78 FPS on an NVIDIA T4 GPU. When pretrained on Objects365, D-FINE-L / X attains 57.1% / 59.3% AP, surpassing all existing real-time detectors. Furthermore, our method significantly enhances the performance of a wide range of DETR models by up to 5.3% AP with negligible extra parameters and training costs. Our code and models: https://github.com/Peterande/D-FINE.

Message passing-based graph neural networks (GNNs) have achieved great success in many real-world applications. For a sampled mini-batch of target nodes, the message passing process is divided into two parts: message passing between nodes within the batch (MP-IB) and message passing from nodes outside the batch to those within it (MP-OB). However, MP-OB recursively relies on higher-order out-of-batch neighbors, leading to an exponentially growing computational cost with respect to the number of layers. Due to the neighbor explosion, the whole message passing stores most nodes and edges on the GPU such that many GNNs are infeasible to large-scale graphs. To address this challenge, we propose an accurate and fast mini-batch approach for large graph transductive learning, namely topological compensation (TOP), which obtains the outputs of the whole message passing solely through MP-IB, without the costly MP-OB. The major pillar of TOP is a novel concept of message invariance, which defines message-invariant transformations to convert costly MP-OB into fast MP-IB. This ensures that the modified MP-IB has the same output as the whole message passing. Experiments demonstrate that TOP is significantly faster than existing mini-batch methods by order of magnitude on vast graphs (millions of nodes and billions of edges) with limited accuracy degradation.

Optimizing computing circuits such as multipliers and adders is a fundamental challenge in modern integrated circuit design. Recent efforts propose formulating this optimization problem as a reinforcement learning (RL) proxy task, offering a promising approach to search high-speed and area-efficient circuit design solutions. However, we show that the RL-based formulation (proxy task) converges to a local optimal design solution (original task) due to the deceptive reward signals and incrementally localized actions in the RL-based formulation. To address this challenge, we propose a novel model-based circuit genetic evolution (MUTE) framework, which reformulates the problem as a genetic evolution process by proposing a grid-based genetic representation of design solutions. This novel formulation avoids misleading rewards by evaluating and improving generated solutions using the true objective value rather than proxy rewards. To promote globally diverse exploration, MUTE proposes a multi-granularity genetic crossover operator that recombines design substructures at varying column ranges between two grid-based genetic solutions. To the best of our knowledge, MUTE is the first to reformulate the problem as a circuit genetic evolution process, which enables effectively searching for global optimal design solutions. We evaluate MUTE on several fundamental computing circuits, including multipliers, adders, and multiply-accumulate circuits. Experiments on these circuits demonstrate that MUTE significantly Pareto-dominates state-of-the-art approaches in terms of both area and delay. Moreover, experiments demonstrate that circuits designed by MUTE well generalize to large-scale computation-intensive circuits as well.

Branch-and-bound methods are pivotal in solving Mixed Integer Linear Programming (MILP), where the challenge of node selection arises, necessitating the prioritization of different regions of the space for subsequent exploration. While machine learning techniques have been proposed to address this, two crucial problems concerning \textbf{(P1)} how to sufficiently extract features from the branch-and-bound tree, and \textbf{(P2)} how to assess the node quality comprehensively based on the features remain open. To tackle these challenges, we propose to tackle the node selection problem employing a novel Tripartite graph representation and Reinforcement learning with a Graph Neural Network model (TRGNN). The tripartite graph is theoretically proved to encompass sufficient information for tree representation in information theory. We learn node selection via reinforcement learning for learning delay rewards and give more comprehensive node metrics. Experiments show that TRGNN significantly improves the efficiency of solving MILPs compared to human-designed and learning-based node selection methods on both synthetic and large-scale real-world MILPs. Moreover, experiments demonstrate that TRGNN well generalizes to MILPs that are significantly larger than those seen during training.

Joint optimization, which jointly optimizes compression and machine vision algorithms, is widely regarded as an effective strategy for enhancing compression performance in the field of coding for machines. However, existing joint optimization methods usually incorporate a semantics parsing module at the end of the pipeline, raising a critical question: Does the performance improvement stem from the joint optimization itself, or is it primarily driven by the tailed semantics parsing module? To address this, we disentangle the tailed semantics parsing module from the joint optimization pipeline by leveraging the simplicity of the person re-identification task, where semantics parsing involves deterministic feature matching rather than a learned neural network. First, we propose a separate optimization pipeline and two joint optimization pipelines to systematically investigate the effectiveness of joint optimization. Our findings reveal that joint optimization alone does not necessarily guarantee performance improvement. Second, we evaluate the influence of the tailed semantics parsing module by equipping it with varying capabilities, demonstrating that higher parsing capability directly correlates with better machine vision performance. These findings underscore the pivotal role of tailed semantics parsing in enhancing machine vision performance and challenge the assumption that joint optimization alone drives improvement. This work offers new insights for designing effective coding methods, emphasizing the interplay between optimization strategies and tailed semantics parsing.

Recent advancements have significantly enhanced the performance of large language models (LLMs) in tackling complex reasoning tasks, achieving notable success in domains like mathematical and logical reasoning. However, these methods encounter challenges with complex planning tasks, primarily due to extended reasoning steps, diverse constraints, and the challenge of handling multiple distinct sub-tasks. To address these challenges, we propose HyperTree Planning (HTP), a novel reasoning paradigm that constructs hypertree-structured planning outlines for effective planning. The hypertree structure enables LLMs to engage in hierarchical thinking by flexibly employing the divide-and-conquer strategy, effectively breaking down intricate reasoning steps, accommodating diverse constraints, and managing multiple distinct sub-tasks in a well-organized manner. We further introduce an autonomous planning framework that completes the planning process by iteratively refining and expanding the hypertree-structured planning outlines. Experiments demonstrate the effectiveness of HTP, achieving state-of-the-art accuracy on the TravelPlanner benchmark with Gemini-1.5-Pro, resulting in a 3.6$\times$ performance improvement over o1-preview.

The prevalent noise in the preference data unavoidably poses significant challenges to the preference alignment of large language models (LLMs). Existing efforts for this problem either marginally alleviate the impact of noise without noise reduction, or rely on external LLMs that incur substantial computational costs. To address these challenges, we propose **RO**bust **P**reference **O**ptimization (**ROPO**), an iterative alignment approach that integrates *noise-tolerance* and *noise filtering* without the aid of external models. Specifically, ROPO first formulates the training process with adaptive noise reduction as an optimization problem, which can be efficiently solved in an iterative paradigm. Then, to equip this solving process with noise-tolerance and noise-identification capabilities, we derive a robust loss that suppresses the gradients from samples with high uncertainty. We demonstrate both empirically and theoretically that the derived loss is key to the noise-tolerance and effective filtering of noisy samples. The derived loss further inspires a robustness-guided rejection sampling technique to compensate for the potential important information in discarded queries. Extensive experiments on several widely-used datasets and model architectures demonstrate that ROPO significantly outperforms all baselines under **four** practical noise settings and the random symmetric noise, with its advantage increasing as the noise rate increases.

Tree-search-based reasoning methods have significantly enhanced the reasoning capability of large language models (LLMs) by facilitating the exploration of multiple intermediate reasoning steps, i.e., thoughts. However, these methods suffer from substantial inference latency, as they have to generate numerous reasoning thoughts, severely limiting LLM applicability. To address this challenge, we propose a novel Speculative Search (SpecSearch) framework that significantly accelerates LLM reasoning by optimizing thought generation. Specifically, SpecSearch utilizes a small model to strategically collaborate with a large model at both thought and token levels, efficiently generating high-quality reasoning thoughts. The major pillar of SpecSearch is a novel quality-preserving rejection mechanism, which effectively filters out thoughts whose quality falls below that of the large model's outputs. Moreover, we show that SpecSearch preserves comparable reasoning quality to the large model. Experiments on both the Qwen and Llama models demonstrate that SpecSearch significantly outperforms state-of-the-art approaches, achieving up to 2.12$\times$ speedup with comparable reasoning quality.

A series of pre-trained models have demonstrated promising results in point cloud understanding tasks and are widely applied to downstream tasks through fine-tuning. However, full fine-tuning leads to the forgetting of pretrained knowledge and substantial storage costs on edge devices. To address these issues, Parameter-Efficient Transfer Learning (PETL) methods have been proposed. According to our analysis, we find that existing 3D PETL methods cannot adequately align with semantic relationships of features required by downstream tasks, resulting in suboptimal performance. To ensure parameter efficiency while introducing rich semantic cues, we propose a novel fine-tuning paradigm for 3D pre-trained models. We utilize frozen 2D pre-trained models to provide vision semantic prompts and design a new Hybrid Attention Adapter to efficiently fuse 2D semantic cues into 3D representations with minimal trainable parameters(1.8M). Extensive experiments conducted on datasets including ScanObjectNN, ModelNet40, and ShapeNetPart demonstrate the effectiveness of our proposed paradigm. In particular, our method achieves 95.6% accuracy on ModelNet40 and attains 90.09% performance on the most challenging classification split ScanObjectNN(PB-T50-RS).

Many real-world applications, such as logistics, routing, scheduling, and production planning, involve dynamic systems that require continuous updates to solutions for new Mixed Integer Linear Programming (MILP) problems. These systems often require rapid updates to their solutions to accommodate slight modifications in constraints or objectives introduced by evolving conditions. While reoptimization techniques have been explored for Linear Programming (LP) and certain specific MILP problems, their effectiveness in addressing general MILP is limited. In this work, we propose a two-stage reoptimization framework for efficiently identifying high-quality feasible solutions. Specifically, we first utilize the historical solving process information to predict a high confidence solution space for modified MILPs, which is likely to contain high-quality solutions. Building on the prediction results, we fix a part of variables within the predicted intervals and apply the Thompson Sampling algorithm to determine which variables to fix. This is done by updating the Beta distributions based on the solutions obtained from the solver. Extensive experiments across nine reoptimization datasets show that our VP-OR outperforms the state-of-the-art methods, achieving higher-quality solutions under strict time limits.

Large-scale sparse linear programming (LP) underpins critical applications in logistics, manufacturing, and scientific computing. However, conventional CPU-only solvers often fail to meet real-time performance demands as problem sizes reach millions of variables and constraints. This paper presents GCP-LP, a GPU-CPU collaborative framework that accelerates the COIN-OR Linear Programming solver by offloading key computational bottlenecks-including sparse matrixvector multiplication (SpMV), pivot selection, and Cholesky factorization-to GPUs while retaining sequential control flow on the CPU. By combining asynchronous data transfer, GPUoptimized memory layouts, and adaptive load balancing, GCPLP achieves up to 10% module-level and 5% overall runtime reductions on NETLIB and Mittelmann benchmarks.

Optimization problems are ubiquitous across various domains, such as resource scheduling, production planning, and sales management. Traditionally, they are modeled manually, leading to inefficiencies due to difficulties in communication and collaboration between modeling and domain experts. The emergence of Large Language Models (LLMs) has made automated modeling possible. However, real-world applications are often large-scale and have numerous variables and constraints, limiting the applicability of existing methods. To address this, we propose PaMOP, a novel modeling framework based on LLMs, to model optimization problems automatically, given only natural language descriptions. Specifically, we extract and partition the problems using a tree structure, guiding the LLMs to model each set of constraints with self-augmented prompts, thus reducing the demands on the LLM's capabilities of large contents. The mathematical model is then iteratively corrected and validated through our correction procedures. The experiments demonstrate that our method improves performance on the common benchmark dataset NLP4LP, achieving an accuracy of 62.3% and a code executability rate of 86.8% when tested on GPT-4. Additionally, we demonstrate the effectiveness of our PaMOP in handling large real-world problems.

Key-Value (KV) cache eviction---which retains the KV pairs of the most important tokens while discarding less important ones---is a critical technique for optimizing both memory usage and inference latency in large language models (LLMs). However, existing approaches often rely on simple heuristics---such as attention weights---to measure token importance, overlooking the spatial relationships between token value states in the vector space. This often leads to suboptimal token selections and thus performance degradation. To tackle this problem, we propose a novel method, namely **AnDPro** (**An**chor **D**irection **Pro**jection), which introduces a projection-based scoring function to more accurately measure token importance. Specifically, AnDPro operates in the space of value vectors and leverages the projections of these vectors onto an *``Anchor Direction''*---the direction of the pre-eviction output---to measure token importance and guide more accurate token selection. Experiments on $16$ datasets from the LongBench benchmark demonstrate that AnDPro can maintain $96.07\\%$ of the full cache accuracy using only $3.44\\%$ KV cache budget, reducing KV cache budget size by $46.0\\%$ without compromising quality compared to previous state-of-the-arts.

While recent multimodal large language models (MLLMs) have advanced automated ECG interpretation, they still face two key limitations: (1) insufficient multimodal synergy between ECG time series and ECG images, and (2) limited explainability in linking diagnoses to granular waveform evidence. We introduce GEM, the first MLLM unifying ECG time series, 12-lead ECG images and text for grounded and clinician-aligned ECG interpretation. GEM enables feature-grounded analysis, evidence-driven reasoning, and a clinician-like diagnostic process through three core innovations: a dual-encoder framework extracting complementary time series and image features, cross-modal alignment for effective multimodal understanding, and knowledge-guided instruction data generation for generating high-granularity grounding data (ECG-Grounding) linking diagnoses to measurable parameters ($e.g.$, QRS/PR Intervals). Additionally, we propose the Grounded ECG Understanding task, a clinically motivated benchmark designed to comprehensively assess the MLLM's capability in grounded ECG understanding. Experimental results on both existing and our proposed benchmarks show GEM significantly improves predictive performance (CSN $7.4\%$ $\uparrow$), explainability ($22.7\%$ $\uparrow$), and grounding ($25.3\%$ $\uparrow$), making it a promising approach for real-world clinical applications. Codes, model, and data are available at https://github.com/lanxiang1017/GEM.

Chip placement is a critical step in the Electronic Design Automation (EDA) workflow, which aims to arrange chip modules on the canvas to optimize the performance, power, and area (PPA) metrics of final designs.Recent advances show great potential of AI-based algorithms in chip placement.However, due to the lengthy EDA workflow, evaluations of these algorithms often focus on intermediate surrogate metrics, which are computationally efficient but often misalign with the final end-to-end performance (i.e., the final design PPA).To address this challenge, we propose to build ChiPBench, a comprehensive benchmark specifically designed to evaluate the effectiveness of AI-based algorithms in final design PPA metrics.Specifically, we generate a diverse evaluation dataset from $20$ circuits across various domains, such as CPUs, GPUs, and NPUs.We then evaluate six state-of-the-art AI-based chip placement algorithms on the dataset and conduct a thorough analysis of their placement behavior.Extensive experiments show that AI-based chip placement algorithms produce unsatisfactory final PPA results, highlighting the significant influence of often-overlooked factors like regularity and dataflow.We believe ChiPBench will effectively bridge the gap between academia and industry.

Arithmetic circuit optimization remains a fundamental challenge in modern integrated circuit design. Recent advances have cast this problem within the Learning to Optimize (L2O) paradigm, where intelligent agents autonomously explore high-performance design spaces with encouraging results. However, existing approaches predominantly target coarse-grained architectural configurations, while the crucial interconnect optimization stage is often relegated to oversimplified proxy models or a heuristic approach. This disconnect undermines design quality, leading to suboptimal solutions in the circuit topology search space. To bridge this gap, we present **Arith-DAS**, a **D**ifferentiable **A**rchitecture **S**earch framework for **Arith**metic circuits. To the best of our knowledge, **Arith-DAS** is the first to formulate interconnect optimization within arithmetic circuits as a differentiable edge prediction problem over a multi-relational directed acyclic graph, enabling fine-grained, proxy-free optimization at the interconnection level. We evaluate **Arith-DAS** on a suite of representative arithmetic circuits, including multipliers and multiply-accumulate units. Experiments show substantial improvements over state-of-the-art L2O and conventional methods, achieving up to $\textbf{27.05}$% gain in hypervolume of area-delay Pareto front, a standard metric for evaluating multi-objective optimization performance. Moreover, integrating our optimized arithmetic units into large-scale AI accelerators yields up to $\textbf{6.59}$% delay reduction, demonstrating both scalability and real-world applicability.

Optimization plays a central role in Operations Research (OR) and numerous industrial applications, yet automating the end-to-end process of translating natural language descriptions into executable optimization programs remains a formidable challenge. While recent efforts have applied Large Language Models (LLMs) to this task, existing approaches are hindered by high inference costs, limited robustness across domains, and weak verification mechanisms. In this work, we propose MURKA, a reinforcement learning and knowledge distillation-based framework that enhances LLM-driven optimization modeling via collaborative agent alignment. MURKA orchestrates three specialized agents---Extractor, Solver, and Checker---to achieve accurate problem understanding, robust formulation, and verifiable execution. The Extractor is trained using group relative policy optimization with a composite reward function that incorporates semantic correctness and execution fidelity. The Solver benefits from knowledge distillation from a powerful teacher model, yielding structurally valid and executable formulations in AMPL. The Checker iteratively verifies solution correctness via solver feedback. We validate MURKA's generalizability through extensive experiments across diverse OR benchmarks, demonstrating its robustness and scalability. Experimental results on eight diverse OR benchmarks, including NLP4LP, ComplexOR, and NL4Opt, demonstrate that MURKA, built on the LLaMa3-8B backbone, achieves a 5.9\% absolute improvement in solution accuracy and a 5.1\% increase in execution success rate compared to leading baselines. These results establish MURKA as an effective and scalable paradigm for LLM-driven optimization, with strong potential for deployment in real-world OR applications.

Cutting planes (cuts) are essential for solving mixed-integer linear programming (MILP) problems, as they tighten the feasible solution space and accelerate the solving process. Modern MILP solvers offer diverse cutting plane separators to generate cuts, enabling users to leverage their potential complementary strengths to tackle problems with different structures. Recent machine learning approaches learn to configure separators based on problem-specific features, selecting effective separators and deactivating ineffective ones to save unnecessary computing time. However, they ignore the dynamics of separator efficacy at different stages of cut generation and struggle to adapt the configurations for the evolving problems after multiple rounds of cut generation. To address this challenge, we propose a novel **dyn**amic **sep**arator configuration (**DynSep**) method that models separator configuration in different rounds as a reinforcement learning task, making decisions based on an incremental triplet graph updated by iteratively added cuts. Specifically, we tokenize the incremental subgraphs and utilize a decoder-only Transformer as our policy to autoregressively predict when to halt separation and which separators to activate at each round. Evaluated on synthetic and large-scale real-world MILP problems, DynSep speeds up average solving time by 64% on easy and medium datasets, and reduces primal-dual gap integral within the given time limit by 16% on hard datasets. Moreover, experiments demonstrate that DynSep well generalizes to MILP instances of significantly larger sizes than those seen during training.

As one of the most critical components in modern LP solvers, presolve in linear programming (LP) employs a rich set of presolvers to remove different types of redundancy in input problems by equivalent transformations. We found from extensive experiments that the presolve routine-that is, the method determining (P1) which presolvers to select, (P2) in what order to execute, and (P3) when to stop-significantly impacts the efficiency of solving LPs. However, designing high-quality presolve routines is highly challenging due to the enormous search space, and further optimizing the routines on different tasks for high performance demands extensive domain knowledge and manual tuning. To tackle this problem, we propose the first learning based framework-that is, reinforcement learning for presolve (RL4Presolve)-to learn high-quality presolve routines. An appealing feature is that we employ a novel adaptive action sequence that learns complex routines efficiently by generating combinations of presolvers automatically at each step. Extensive experiments demonstrate that RL4Presolve achieves significant improvement (up to roughly 90% ) in the efficiency of solving LPs. Furthermore, we extract routines from learned policies for simple and efficient deployment without GPU resources to Huawei's supply chain, where extensive manual tuning for each separate task was required previously due to the high economic value.

Temporal action detection aims to predict temporal boundaries and category labels of actions in untrimmed videos. In the past years, many weakly supervised temporal action detection methods have been proposed to relieve the annotation cost of fully supervised methods. Due to the discrepancy between action localization and action classification, the two-branch structure is widely adopted by existing weakly supervised methods, where the classification branch is used to predict category-wise score and the localization branch is used to predict foreground score for each segment. Under the weakly supervised setting, the model training is mainly guided by the video-level or sparse segment-level annotations. As a result, the classification branch tends to focus on the most discriminative segments while ignore less discriminative ones so as to minimize the classification cost, and the localization branch may assign high foreground scores for some negative segments. This phenomenon can severely damage the action detection performance, because the foreground scores and classification scores are combined together in the testing stage for action detection. To deal with this problem, several methods have been proposed to encourage the consistency between the classification branch and localization branch. However, these methods only consider the video-level or segment-level consistency, without considering the relation among different segments to be consistent. In this paper, we propose a Cross-Task Relation-Aware Consistency (CRC) strategy for weakly supervised temporal action detection, including an intra-video consistency module and an inter-video consistency module. The intra-video consistency module can well guarantee the relationship among segments from the same video to be consistent, and the inter-video consistency module guarantees the relationship among segments from different videos to be consistent. These two modules are complementary to each other by combining both intra-video and inter-video consistency. Experimental results show that the proposed CRC strategy can consistently improve the performance of existing weakly supervised methods, including click-level supervised methods (e.g., LACP Lee et al., 2021), video-level supervised methods (e.g., DELU Chen et al., 2022) and unsupervised methods (e.g., BaS-Net Lee et al., 2020), verifying the generality and effectiveness of the proposed method.

Semi-supervised 3D object detection from point cloud aims to train a detector with a small number of labeled data and a large number of unlabeled data. Among existing methods, the pseudo-label based methods have achieved superior performance, and the core lies in how to select high-quality pseudo-labels with the designed quality evaluation criterion. Despite the success of these methods, they all consider the localization and classification quality estimation from a global perspective. For localization quality, they use a global score threshold to filter out low-quality pseudo-labels and assign equal importance to each side during training, ignoring the fact that sides with different localization quality should not be treat equally. Besides, a large number of pseudo-labels are discarded due to the high global threshold, which may also contain some correctly predicted sides that are helpful for model training. For the classification quality, they usually combine the objectness score and classification confidence score to filter out pseudo-labels. The main focus of them is designing effective classification confidence evaluation metrics, neglecting the importance of predicting better objectness score. In this paper, we propose SA3Det++, a side-aware quality estimation method for semi-supervised object detection, which consists of a probabilistic side localization strategy, a side-aware quality estimation strategy, and a soft pseudo-label selection strategy. Extensive results demonstrate that the proposed method consistently outperforms the baseline methods under different scenes and evaluation criterions.

Click-based interactive segmentation is the most concise and widely used data labeling method. While existing interactive segmentation methods excel in handling simple targets, they encounter challenges in obtaining high-quality masks from some complex scenes, even with a large number of clicks. Also, the cost of retraining the model from scratch for special scenarios is unacceptably high. To address these issues, we propose ClickAdapter, a simple yet powerful interactive segmentation model adapter without the need for no pre-training. Through introducing a small number of additional parameters and computations, the adapter module effectively enhanced the ability of interactive segmentation models to obtain high-quality prediction with limited clicks. Specifically, we incorporate a detail extractor that aims to extract spatial correlations and local detail features of images. These fine-grained data are then integrated into a model with our adapter to generate segmentation masks with sharp and precise edges. During the training process, only the parameters of our adapter are learnable, thereby reducing the training cost. Features in special scenarios can also be infused more efficiently. To verify the efficiency and performance advantages of the proposed method, a series of experiments on a wide range of benchmarks were conducted, demonstrating that the proposed algorithm achieved cutting-edge performance compared to current state-of-the-art (SOTA) methods.

Few-shot video object segmentation (FSVOS) aims to achieve accurate segmentation of novel objects in given video sequences, where the target objects are specified by limited annotated images as support. Most previous top-performing methods adopt the support-query semantic correlation learning paradigm or the intra-query temporal correlation learning paradigm. Nevertheless, they either fail to model temporal consistency across frames, resulting in inconsecutive segmentation, or lose diverse support object information, leading to incomplete segmentation. Therefore, we argue that it is more desirable to achieve both correlations in a collaborative manner. In this work, we delve into the issues present in the combination of few-shot image segmentation methods and video object segmentation methods and propose a dedicated Collaborative Correlation Network (CoCoNet) to address these problems, including a pixel correlation calibration module and a temporal correlation mining module. The proposed CoCoNet enjoys several merits. First, the pixel correlation calibration module aims to mitigate the noise issue in support-query correlation by integrating the affinity learning strategy and the prototype learning strategy. Specifically, we employ Optimal Transport to enrich pixel correlation with contextual information, thereby reducing intra-class differences between support and query. Second, the temporal correlation mining module is responsible for alleviating the issue of uncertainty in the initial frame and establishing reliable guidance for subsequent frames of the query video. With the collaboration of these two modules, our CoCoNet can effectively establish support-query and temporal correlation simultaneously and achieve accurate FSVOS. Extensive experimental results on two challenging benchmarks demonstrate that our method performs favorably against state-of-the-art FSVOS methods.

Image registration plays a crucial role in histological image analysis, encompassing tasks like multi-modality fusion and disease grading. Traditional registration methods optimize objective functions for each image pair, yielding reliable accuracy but demanding heavy inference burdens. Recently, learning-based registration methods utilize networks to learn the optimization process during training and apply a one-step forward process during testing. While these methods offer promising registration performance with reduced inference time, they remain sensitive to appearance variances and local structure changes commonly encountered in histological image registration scenarios. In this paper, for the first time, we propose a novel test-time adaptation method for histological image registration, aiming to improve the generalization ability of learning-based methods. Specifically, we design two operations, style guidance and shape guidance, for the test-time adaptation process. The former leverages style representations encoded by feature statistics to address the issue of appearance variances, while the latter incorporates shape representations encoded by HOG features to improve registration accuracy in regions with structural changes. Furthermore, we consider the continuity of the model during the test-time adaptation process. Different from the previous methods initialized by a given trained model, we introduce a smoothing strategy to leverage historical models for better generalization. We conduct experiments with several representative learning-based backbones on the public histological dataset, demonstrating the superior registration performance of our test-time adaptation method.

The current neuron reconstruction pipeline for electron microscopy (EM) data usually includes automatic image segmentation followed by extensive human expert proofreading. In this work, we aim to reduce human workload by predicting connectivity between over-segmented neuron pieces, taking both microscopy image and 3D morphology features into account, similar to human proofreading workflow. To this end, we first construct a dataset, named FlyTracing, that contains millions of pairwise connections of segments expanding the whole fly brain, which is three orders of magnitude larger than existing datasets for neuron segment connection. To learn sophisticated biological imaging features from the connectivity annotations, we propose a novel connectivity-aware contrastive learning method to generate dense volumetric EM image embedding. The learned embeddings can be easily incorporated with any point or voxel-based morphological representations for automatic neuron tracing. Extensive comparisons of different combination schemes of image and morphological representation in identifying split errors across the whole fly brain demonstrate the superiority of the proposed approach, especially for the locations that contain severe imaging artifacts, such as section missing and misalignment. The dataset and code are available at https://github.com/Levishery/Flywire-Neuron-Tracing.

Automatic mitochondrial segmentation enjoys great popularity with the development of deep learning. However, the coarse prediction raised by the presence of regular 3D grids in previous methods regardless of 3D CNN or the vision transformers suggest a possibly sub-optimal feature arrangement. To mitigate this limitation, we attempt to interpret the 3D EM image stacks as a set of interrelated 3D fragments for a better solution. However, it is non-trivial to model the 3D fragments without introducing excessive computational overhead. In this paper, we design a coherent fragment vision transformer (FragViT) combined with affinity learning to manipulate features on 3D fragments yet explore mutual relationships to model fragment-wise context, enjoying locality prior without sacrificing global reception. The proposed FragViT includes a fragment encoder and a hierarchical fragment aggregation module. The fragment encoder is equipped with affinity heads to transform the tokens into fragments with homogeneous semantics, and the multi-layer self-attention is used to explicitly learn inter-fragment relations with long-range dependencies. The hierarchical fragment aggregation module is responsible for hierarchically aggregating fragment-wise prediction back to the final voxel-wise prediction in a progressive manner. Extensive experimental results on the challenging MitoEM, Lucchi, and AC3/AC4 benchmarks demonstrate the effectiveness of the proposed method.

Video semantic segmentation has achieved conspicuous achievements attributed to the development of deep learning, but suffers from labor-intensive annotated training data gathering. To alleviate the data-hunger issue, domain adaptation approaches are developed in the hope of adapting the model trained on the labeled synthetic videos to the real videos in the absence of annotations. By analyzing the dominant paradigm consistency regularization in the domain adaptation task, we find that the bottlenecks exist in previous methods from the perspective of pseudo-labels. To take full advantage of the information contained in the pseudo-labels and empower more effective supervision signals, we propose a coherent PAT network including a target domain focalizer and relation-aware temporal consistency. The proposed PAT network enjoys several merits. First, the target domain focalizer is responsible for paying attention to the target domain, and increasing the accessibility of pseudo-labels in consistency training. Second, the relation-aware temporal consistency aims at modeling the inter-class consistent relationship across frames to equip the model with effective supervision signals. Extensive experimental results on two challenging benchmarks demonstrate that our method performs favorably against state-of-the-art domain adaptive video semantic segmentation methods.

Chain-of-Thought prompting has improved reasoning capability of large language models (LLM). However, it still is challenging to guarantee the effectiveness and stability for questions requiring complicated reasoning. Recently, Plan-and-Solve prompting enhances the reasoning capability for complex questions by planning the solution steps firstly and then solving them step by step, but it suffers the difficulty to represent and execute the problem-solving logic of complex questions. To deal with these challenges, in this work, we propose a novel Plan-and-Solve prompting method based on Question Decomposition Meaning Representation (QDMR). Specifically, this method first allows the LLM to generate a QDMR graph to represent the problem-solving logic, which is a directed acyclic graph composed of sub-questions. Then, the LLM generates a specific solving process based on the QDMR graph. When solving each sub-question, it can locate the preceding sub-questions and their answers according to the QDMR graph, and then utilize this information for solution. Compared with existing Plan-and-Solve prompting techniques, our method can not only represent the problem-solving logic of complicated questions more accurately with the aid of QDMR graph, but also deliver the dependence information accurately for different solution steps according to the QDMR graph. In addition, with the supervised fine-tuning on the Allen Institute dataset, the decomposing capability of LLM for complicated questions can be considerably enhanced. Extensive experiments show that our method has achieve a great significance in arithmetic reasoning and commonsense reasoning task by comparing the classical Chain-of-Thought prompting and Plan-and-Solve prompting techniques, and the improvements achieved are even greater for problems with more reasoning steps.

The key lie in semi-supervised semantic segmentation is how to fully exploit substantial unlabeled data to im-prove the model's generalization performance by resorting to constructing effective supervision signals. Most methods tend to directly apply contrastive learning to seek additional supervision to complement independent regular pixel-wise consistency regularization. However, these methods tend not to be preferred ascribed to their complicated designs, heavy memory footprints and susceptibility to confirmation bias. In this paper, we analyze the bottlenecks exist in con-trastive learning-based methods and offer a fresh perspective on inter-pixel correlations to construct more safe and effective supervision signals, which is in line with the nature of semantic segmentation. To this end, we develop a coherent RankMatch network, including the construction of representative agents to model inter-pixel correlation beyond regular individual pixel-wise consistency, and fur-ther unlock the potential of agents by modeling inter-agent relationships in pursuit of rank-aware correlation consis-tency. Extensive experimental results on multiple bench-marks, including mitochondria segmentation, demonstrate that RankMatch performs favorably against state-of-the-art methods. Particularly in the low-data regimes, RankMatch achieves significant improvements.

As the mainstream technique for capturing images of biological specimens at nanometer resolution, electron microscopy (EM) is extremely time-consuming for scanning wide field-of-view (FOV) specimens. In this paper, we investigate a challenging task of large-factor EM image super-resolution (EMSR), which holds great promise for reducing scanning time, relaxing acquisition conditions, and expanding imaging FOV. By exploiting the repetitive structures and volumetric coherence of EM images, we propose the first generative learning-based framework for large-factor EMSR. Specifically, motivated by the predictability ofrepetitive structures and textures in EM images, we first learn a discrete codebook in the latent space to represent highresolution (HR) cell-specific priors and a latent vector indexer to map low-resolution (LR) EM images to their corresponding latent vectors in a generative manner. By incorporating the generative cell-specific priors from HR EM images through a multi-scale prior fusion module, we then deploy multi-image feature alignment and fusion to further exploit the inter-section coherence in the volumetric EM data. Extensive experiments demonstrate that our proposed framework outperforms advanced single-image and video super-resolution methods for 8× and 16× EMSR (i.e., with 64 times and 256 times less data acquired, respectively), achieving superior visual reconstruction quality and down-stream segmentation accuracy on benchmark EM datasets. Code is available at https://github.com/jtshou/GPEMSR.

In this work, we investigate the joint optimization strategy in the scenario of video coding for machines (VCM). We formulated two kinds of joint optimization strategies, Opt_JA and Opt_JH , and compared them with the separate optimization strategy Opt_S. The three optimization strategies are illustrated in Fig. 1 . In Opt_S , we separately train the feature compression network with mean squared error (MSE). In Opt_JA , we optimize all modules jointly toward the person re-identification task. In Opt_JH , only the aggregation module and feature compression module are jointly optimized. The feature compression consists of two fully-connected (FC) layers and two batch normalization (BN) layers. Specifically, we set five compression ratios (CR): 256, 128, 64, 32, and 16.

Machine learning (ML) has been shown to successfully accelerate solving NP-hard combinatorial optimization (CO) problems under the branch and bound framework. However, the high training and inference cost and limited interpretability of ML approaches severely limit their wide application to modern exact CO solvers. In contrast, human-designed policies---though widely integrated in modern CO solvers due to their compactness and reliability---can not capture data-driven patterns for higher performance. To combine the advantages of the two paradigms, we propose the first symbolic discovery framework---namely, deep symbolic discovery for exact combinatorial optimization solver (Symb4CO)---to learn high-performance symbolic policies on the branching task. Specifically, we show the potential existence of small symbolic policies empirically, employ a large neural network to search in the high-dimensional discrete space, and compile the learned symbolic policies directly for fast deployment. Experiments show that the Symb4CO learned purely CPU-based policies consistently achieve *comparable* performance to previous GPU-based state-of-the-art approaches. Furthermore, the appealing features of Symb4CO include its high training (*ten training instances*) and inference (*one CPU core*) efficiency and good interpretability (*one-line expressions*), making it simple and reliable for deployment. The results show encouraging potential for the *wide* deployment of ML to modern CO solvers.

Learning neural operators for solving partial differential equations (PDEs) has attracted great attention due to its high inference efficiency. However, training such operators requires generating a substantial amount of labeled data, i.e., PDE problems together with their solutions. The data generation process is exceptionally time-consuming, as it involves solving numerous systems of linear equations to obtain numerical solutions to the PDEs. Many existing methods solve these systems independently without considering their inherent similarities, resulting in extremely redundant computations. To tackle this problem, we propose a novel method, namely **S**orting **K**rylov **R**ecycling (**SKR**), to boost the efficiency of solving these systems, thus significantly accelerating data generation for neural operators training. To the best of our knowledge, SKR is the first attempt to address the time-consuming nature of data generation for learning neural operators. The working horse of SKR is Krylov subspace recycling, a powerful technique for solving a series of interrelated systems by leveraging their inherent similarities. Specifically, SKR employs a sorting algorithm to arrange these systems in a sequence, where adjacent systems exhibit high similarities. Then it equips a solver with Krylov subspace recycling to solve the systems sequentially instead of independently, thus effectively enhancing the solving efficiency. Both theoretical analysis and extensive experiments demonstrate that SKR can significantly accelerate neural operator data generation, achieving a remarkable speedup of up to 13.9 times.

Macro placement is a crucial step in modern chip design, and reinforcement learning (RL) has recently emerged as a promising technique for improving the placement quality. However, existing RL-based techniques are hindered by their low sample efficiency, requiring numerous online rollouts or substantial offline expert data to achieve bootstrap, which are often impractical in industrial scenarios. To address this challenge, we propose a novel sample-efficient framework, namely **EfficientPlace**, for fast macro placement. EfficientPlace integrates a global tree search algorithm to strategically direct the optimization process, as well as a RL agent for local policy learning to advance the tree search. Experiments on commonly used benchmarks demonstrate that EfficientPlace achieves remarkable placement quality within a short timeframe, outperforming recent state-of-the-art approaches.

Machine learning (ML) approaches have been successfully applied to accelerating exact combinatorial optimization (CO) solvers. However, many of them fail to explain what patterns they have learned that accelerate the CO algorithms due to the black-box nature of ML models like neural networks, and thus they prevent researchers from further understanding the tasks they are interested in. To tackle this problem, we propose the first graph-based algorithm discovery framework—namely, graph symbolic discovery for exact combinatorial optimization solver (GS4CO)—that learns interpretable branching policies directly from the general bipartite graph representation of CO problems. Specifically, we design a unified representation for symbolic policies with graph inputs, and then we employ a Transformer with multiple tree-structural encodings to generate symbolic trees end-to-end, which effectively reduces the cumulative error from iteratively distilling graph neural networks. Experiments show that GS4CO learned interpretable and lightweight policies outperform all the baselines on CPU machines, including both the human-designed and the learning-based. GS4CO shows an encouraging step towards general algorithm discovery on modern CO solvers.

Generation of plausible but incorrect factual information, often termed hallucination, has attracted significant research interest. Retrieval-augmented language model (RALM)---which enhances models with up-to-date knowledge---emerges as a promising method to reduce hallucination. However, existing RALMs may instead exacerbate hallucination when retrieving lengthy contexts. To address this challenge, we propose COFT, a novel **CO**arse-to-**F**ine highligh**T**ing method to focus on different granularity-level key texts, thereby avoiding getting lost in lengthy contexts. Specifically, COFT consists of three components: *recaller*, *scorer*, and *selector*. First, *recaller* applies a knowledge graph to extract potential key entities in a given context. Second, *scorer* measures the importance of each entity by calculating its contextual weight. Finally, *selector* selects high contextual weight entities with a dynamic threshold algorithm and highlights the corresponding paragraphs, sentences, or words in a coarse-to-fine manner. Extensive experiments on knowledge hallucination benchmark demonstrate the effectiveness of COFT, leading to a superior performance over 30% in F1 score metric. Moreover, COFT also exhibits remarkable versatility across various long-form tasks, such as reading comprehension and question answering.

Logic Synthesis (LS) plays a vital role in chip design. A key task in LS is to simplify circuits---modeled by directed acyclic graphs (DAGs)---with functionality-equivalent transformations. To tackle this task, many LS heuristics apply transformations to subgraphs---rooted at each node on an input DAG---sequentially. However, we found that a large number of transformations are ineffective, which makes applying these heuristics highly time-consuming. In particular, we notice that the runtime of the Resub and Mfs2 heuristics often dominates the overall runtime of LS optimization processes. To address this challenge, we propose a novel data-driven LS heuristic paradigm, namely PruneX, to reduce ineffective transformations. The major challenge of developing PruneX is to learn models that well generalize to unseen circuits, i.e., the out-of-distribution (OOD) generalization problem. Thus, the major technical contribution of PruneX is the novel circuit domain generalization framework, which learns domain-invariant representations based on the transformation-invariant domain-knowledge. To the best of our knowledge, PruneX is the first approach to tackle the OOD problem in LS heuristics. We integrate PruneX with the aforementioned Resub and Mfs2 heuristics. Experiments demonstrate that PruneX significantly improves their efficiency while keeping comparable optimization performance on industrial and very large-scale circuits, achieving up to $3.1\times$ faster runtime.

Multiplier design—which aims to explore a large combinatorial design space to simultaneously optimize multiple conflicting objectives—is a fundamental problem in the integrated circuits industry. Although traditional approaches tackle the multi-objective multiplier optimization problem by manually designed heuristics, reinforcement learning (RL) offers a promising approach to discover high-speed and area-efficient multipliers. However, the existing RL-based methods struggle to find Pareto-optimal circuit designs for all possible preferences, i.e., weights over objectives, in a sample-efficient manner. To address this challenge, we propose a novel hierarchical adaptive (HAVE) multi-task reinforcement learning framework. The hierarchical framework consists of a meta-agent to generate diverse multiplier preferences, and an adaptive multi-task agent to collaboratively optimize multipliers conditioned on the dynamic preferences given by the meta-agent. To the best of our knowledge, HAVE is the first to well approximate Pareto-optimal circuit designs for the entire preference space with high sample efficiency. Experiments on multipliers across a wide range of input widths demonstrate that HAVE significantly Pareto-dominates state-of-the-art approaches, achieving up to 28% larger hypervolume. Moreover, experiments demonstrate that multipliers designed by HAVE can well generalize to large-scale computation-intensive circuits.

Offline reinforcement learning (RL) provides a promising approach to avoid costly online interaction with the real environment. However, the performance of offline RL highly depends on the quality of the datasets, which may cause extrapolation error in the learning process. In many robotic applications, an inaccurate simulator is often available. However, the data directly collected from the inaccurate simulator cannot be directly used in offline RL due to the well-known exploration-exploitation dilemma and the dynamic gap between inaccurate simulation and the real environment. To address these issues, we propose a novel approach to combine the offline dataset and the inaccurate simulation data in a better manner. Specifically, we pre-train a generative adversarial network (GAN) model to fit the state distribution of the offline dataset. Given this, we collect data from the inaccurate simulator starting from the distribution provided by the generator and reweight the simulated data using the discriminator. Our experimental results in the D4RL benchmark and a real-world manipulation task confirm that our method can benefit more from both inaccurate simulator and limited offline datasets to achieve better performance than the state-of-the-art methods.

As one of the key aspects of Versatile Video Coding (VVC), the quad-tree with a nested multi-type tree (QTMT) partition structure enhances the rate-distortion (RD) performance but at the cost of extensive computational encoding complexity. To reduce the complexity of QTMT partition in VVC intraframe coding, researchers proposed the partition map-based fast block partitioning algorithm, which achieves advanced encoding time savings and coding efficiency. However, it encounters high inference overhead due to the over-parameterized neural network. To efficiently deploy this algorithm, we first propose a lightweight neural network based on the hierarchical vision transformer that predicts the partition map effectively with restricted computational complexity, thereby reducing the inference complexity uniformly. Next, we introduce token sparsification to select the most informative tokens using a predefined pruning ratio, achieving content-adaptive computation reduction and parallel-friendly inference. Experimental results demonstrate that the proposed method provides a reduction of 98.71% FLOPs with a negligible BDBR increase compared to the original methods. The source code is available at https://github.com/ustc-ivclab/EPM.

Mixed-integer linear programming (MILP) is one of the most popular mathematical formulations with numerous applications. In practice, improving the performance of MILP solvers often requires a large amount of high-quality data, which can be challenging to collect. Researchers thus turn to generation techniques to generate additional MILP instances. However, existing approaches do not take into account specific block structures—which are closely related to the problem formulations—in the constraint coefficient matrices (CCMs) of MILPs. Consequently, they are prone to generate computationally trivial or infeasible instances due to the disruptions of block structures and thus problem formulations. To address this challenge, we propose a novel MILP generation framework, called Block Structure Decomposition (MILP-StuDio), to generate high-quality instances by preserving the block structures. Specifically, MILP-StuDio begins by identifying the blocks in CCMs and decomposing the instances into block units, which serve as the building blocks of MILP instances. We then design three operators to construct new instances by removing, substituting, and appending block units in the original instances, enabling us to generate instances with flexible sizes. An appealing feature of MILP-StuDio is its strong ability to preserve the feasibility and computational hardness of the generated instances. Experiments on the commonly-used benchmarks demonstrate that using instances generated by MILP-StuDio is able to significantly reduce over 10% of the solving time for learning-based solvers.

Logic Synthesis (LS) aims to generate an optimized logic circuit satisfying a given functionality, which generally consists of circuit translation and optimization. It is a challenging and fundamental combinatorial optimization problem in integrated circuit design. Traditional LS approaches rely on manually designed heuristics to tackle the LS task, while machine learning recently offers a promising approach towards next-generation logic synthesis by neural circuit generation and optimization. In this paper, we first revisit the application of differentiable neural architecture search (DNAS) methods to circuit generation and found from extensive experiments that existing DNAS methods struggle to exactly generate circuits, scale poorly to large circuits, and exhibit high sensitivity to hyper-parameters. Then we provide three major insights for these challenges from extensive empirical analysis: 1) DNAS tends to overfit to too many skip-connections, consequently wasting a significant portion of the network's expressive capabilities; 2) DNAS suffers from the structure bias between the network architecture and the circuit inherent structure, leading to inefficient search; 3) the learning difficulty of different input-output examples varies significantly, leading to severely imbalanced learning. To address these challenges in a systematic way, we propose a novel regularized triangle-shaped circuit network generation framework, which leverages our key insights for completely accurate and scalable circuit generation. Furthermore, we propose an evolutionary algorithm assisted by reinforcement learning agent restarting technique for efficient and effective neural circuit optimization. Extensive experiments on four different circuit benchmarks demonstrate that our method can precisely generate circuits with up to 1200 nodes. Moreover, our synthesized circuits significantly outperform the state-of-the-art results from several competitive winners in IWLS 2022 and 2023 competitions.

In-loop filtering (ILF) is a key technology in image/video coding for reducing the artifacts. Recently, neural network-based in-loop filtering methods achieve remarkable coding gains beyond the capability of advanced video coding standards, establishing themselves a promising candidate tool for future standards. However, the utilization of deep neural networks (DNN) brings high computational complexity and raises high demand of dedicated hardware, which is challenging to apply into general use. To address this limitation, we study an efficient in-loop filtering scheme by adopting look-up tables (LUTs). After training a DNN with a predefined reference range for in-loop filtering, we cache the output values of the DNN into a LUT via traversing all possible inputs. In the coding process, the filtered pixel is generated by locating the input pixels (to-be-filtered pixel and reference pixels) and interpolating between the cached values. To further enable larger reference range within the limited LUT storage, we introduce an enhanced indexing mechanism in the filtering process, and a clipping/finetuning mechanism in the training. The proposed method is implemented into the Versatile Video Coding (VVC) reference software, VTM-11.0. Experimental results show that the proposed method, with three different configurations, achieves on average 0.13%∼0.51%, and 0.10% ∼0.39% BD-rate reduction under the all-intra (AI) and random-access (RA) configurations respectively. The proposed method incurs only 1% ∼8% time increase, an additional computation of 0.13 ∼0.93 kMAC/pixel, and 164 ∼1148 KB storage cost for a single model. Our method has explored a new and more practical approach for neural network-based ILF.

Inter prediction is a key technology in video coding to reduce the temporal redundancy. In natural videos, there are usually moving objects with changing velocity, resulting in complex motion fields that are difficult to represent compactly. In Versatile Video Coding (VVC), existing inter prediction methods usually assume uniform speed motion between consecutive frames, which may not well handle the complex motion fields in the real world. To address these issues, we introduce a uniformly accelerated motion model (UAMM) to exploit velocity and acceleration of moving objects between the video frames, and further combine them to assist in the inter prediction methods to handle the motion change in the temporal domain. First, we review the theory of UAMM. Second, we propose UAMM-based parameter derivation and extrapolation schemes in the coding process. Third, we integrate the UAMM into existing inter prediction modes (Merge, MMVD, CIIP) to achieve higher prediction accuracy. The proposed method is implemented into the VVC reference software, VTM version 12.0. Experimental results show that the proposed method achieves up to 0.38% BD-rate reduction compared to the VTM anchor, under the Low-delay P configuration, with a slight increase of time complexity on the encoding/decoding side.

Cutting planes (cuts) play an important role in solving mixed-integer linear programs (MILPs), which formulate many important real-world applications. Cut selection heavily depends on (P1) which cuts to prefer and (P2) how many cuts to select. Although modern MILP solvers tackle (P1)-(P2) by human-designed heuristics, machine learning carries the potential to learn more effective heuristics. However, many existing learning-based methods learn which cuts to prefer, neglecting the importance of learning how many cuts to select. Moreover, we observe that (P3) what order of selected cuts to prefer significantly impacts the efficiency of MILP solvers as well. To address these challenges, we propose a novel hierarchical sequence/set model (HEM) to learn cut selection policies. Specifically, HEM is a bi-level model: (1) a higher-level module that learns how many cuts to select, (2) and a lower-level module-that formulates the cut selection as a sequence/set to sequence learning problem-to learn policies selecting an ordered subset with the cardinality determined by the higher-level module. To the best of our knowledge, HEM is the first data-driven methodology that well tackles (P1)-(P3) simultaneously. Experiments demonstrate that HEM significantly improves the efficiency of solving MILPs on eleven challenging MILP benchmarks, including two Huawei's real problems.

Different from visible cameras which record intensity images frame by frame, the biologically inspired event camera produces a stream of asynchronous and sparse events with much lower latency. In practice, visible cameras can better perceive texture details and slow motion, while event cameras can be free from motion blurs and have a larger dynamic range which enables them to work well under fast motion and low illumination (LI). Therefore, the two sensors can cooperate with each other to achieve more reliable object tracking. In this work, we propose a large-scale Visible-Event benchmark (termed VisEvent) due to the lack of a realistic and scaled dataset for this task. Our dataset consists of 820 video pairs captured under LI, high speed, and background clutter scenarios, and it is divided into a training and a testing subset, each of which contains 500 and 320 videos, respectively. Based on VisEvent, we transform the event flows into event images and construct more than 30 baseline methods by extending current single-modality trackers into dual-modality versions. More importantly, we further build a simple but effective tracking algorithm by proposing a cross-modality transformer, to achieve more effective feature fusion between visible and event data. Extensive experiments on the proposed VisEvent dataset, FE108, COESOT, and two simulated datasets (i.e., OTB-DVS and VOT-DVS), validated the effectiveness of our model. The dataset and source code have been released on: https://github.com/wangxiao5791509/VisEvent_SOT_Benchmark.

Recent learning-based methods demonstrate their strong ability to estimate depth for multi-view stereo reconstruction. However, most of these methods directly extract features via regular or deformable convolutions, and few works consider the alignment of the receptive fields between views while constructing the cost volume. Through analyzing the constraint and inference of previous MVS networks, we find that there are still some shortcomings that hinder the performance. To deal with the above issues, we propose an Epipolar-Guided Multi-View Stereo Network with Interval-Aware Label (EI-MVSNet), which includes an epipolar-guided volume construction module and an interval-aware depth estimation module in a unified architecture for MVS. The proposed EI-MVSNet enjoys several merits. First, in the epipolar-guided volume construction module, we construct cost volume with features from aligned receptive fields between different pairs of reference and source images via epipolar-guided convolutions, which take rotation and scale changes into account. Second, in the interval-aware depth estimation module, we attempt to supervise the cost volume directly and make depth estimation independent of extraneous values by perceiving the upper and lower boundaries, which can achieve fine-grained predictions and enhance the reasoning ability of the network. Extensive experimental results on two standard benchmarks demonstrate that our EI-MVSNet performs favorably against state-of-the-art MVS methods. Specifically, our EI-MVSNet ranks$1_{st}$on both intermediate and advanced subsets of the Tanks and Temples benchmark, which verifies the high precision and strong robustness of our model.

The mainstream of image and sentence matching studies currently focuses on fine-grained alignment of image regions and sentence words. However, these methods miss a crucial fact: the correspondence between images and sentences does not simply come from alignments between individual regions and words but from alignments between the phrases they form respectively. In this work, we propose a novel Decoupled Cross-modal Phrase-Attention network (DCPA) for image-sentence matching by modeling the relationships between textual phrases and visual phrases. Furthermore, we design a novel decoupled manner for training and inferencing, which is able to release the trade-off for bi-directional retrieval, where image-to-sentence matching is executed in textual semantic space and sentence-to-image matching is executed in visual semantic space. Extensive experimental results on Flickr30K and MS-COCO demonstrate that the proposed method outperforms state-of-the-art methods by a large margin, and can compete with some methods introducing external knowledge.

Generalized Zero-Shot Learning (GZSL) aims at recognizing images from both seen and unseen classes by constructing correspondences between visual images and semantic embedding. However, existing methods suffer from a strong bias problem, where unseen images in the target domain tend to be recognized as seen classes in the source domain. To address this issue, we propose a Prototype-augmented Self-supervised Generative Network by integrating self-supervised learning and prototype learning into a feature generating model for GZSL. The proposed model enjoys several advantages. First, we propose a Self-supervised Learning Module to exploit inter-domain relationships, where we introduce anchors as a bridge between seen and unseen categories. In the shared space, we pull the distribution of the target domain away from the source domain and obtain domain-aware features. To our best knowledge, this is the first work to introduce self-supervised learning into GZSL as learning guidance. Second, a Prototype Enhancing Module is proposed to utilize class prototypes to model reliable target domain distribution in finer granularity. In this module, a Prototype Alignment mechanism and a Prototype Dispersion mechanism are combined to guide the generation of better target class features with intra-class compactness and inter-class separability. Extensive experimental results on five standard benchmarks demonstrate that our model performs favorably against state-of-the-art GZSL methods.

Multi-intersection traffic signal control (TSC) is an active research field in multi-agent systems, where traffic signals for each intersection, controlled by an agent, must coordinate to optimize traffic flow. To encourage global coordination, previous work partitions the traffic network into several regions and learns policies for agents in a feudal structure. However, static network partition fails to adapt to dynamic traffic flow, which changes frequently over time. To address this, we propose a novel multi-agent reinforcement learning approach with adaptive network partition. Specifically, we partition the network into several regions according to the dynamic traffic flow over time. To do this, we propose two approaches: one is directly to use graphic neural network (GNN) to generate the network partition, and the other is to use Monte-Carlo tree search (MCTS) to find the best partition with criteria computed by GNN. Then, we design a variant of Qmix using GNN to handle various dimensions of input, given by the dynamic network partition. Finally, we use a feudal hierarchy to manage agents in each partition and promote global cooperation. By doing so, agents are able to adapt to the traffic flow as required in practice. We empirically evaluate our method both in a synthetic traffic grid and real-world traffic networks of three cities, widely used in the literature. The experimental results confirm that our method achieved better performance both in a synthetic traffic grid and real-world traffic networks of three cities, in terms of average travel time and queue length, than several leading TSC baselines.

Compared to many other dense prediction tasks, object detection plays a fundamental role in visual perception and scene understanding. Dense object detection, aiming at localizing objects directly from the feature map, has drawn great attention due to its low cost and high efficiency. Though it has been developed for a long time, the training pipeline of dense object detectors is still compromised to lots of conjunctions. In this paper, we demonstrate the existence of three conjunctions lying in the current paradigm of one-stage detectors: 1) only samples assigned as positive in classification head are used to train the regression head; 2) classification and regression share the same input feature and computational fields defined by the parallel head architecture; and 3) samples distributed in different feature pyramid layers are treated equally when computing the loss. Based on this, we propose Disentangled Dense Object Detector (DDOD), a simple, direct, and efficient framework for 2D detection with strong performance. We derive two DDOD variants (i.e., DR-CNN, and DDETR) following the basic one-stage/two-stage and recently developed transformer-based pipelines. Specifically, we develop three effective disentanglement mechanisms and integrate them into the current state-of-the-art object detectors. Extensive experiments on MS COCO benchmark show that our approach obtains significant enhancements with negligible extra overhead on various detectors. Notably, our best model reaches 55.4 mAP on the COCOtest-devset, achieving new state-of-the-art performance on this competitive benchmark. Additionally, we validate our model on several challenging tasks including small object detection and crowded object detection. The experimental results further prove the superiority of disentanglement on these conjunctions. Code is available athttps://github.com/zehuichen123/DDOD.

The analysis of neuronal morphological data is essential to investigate the neuronal properties and brain mechanisms. The complex morphologies, absence of annotations, and sheer volume of these data pose significant challenges in neuronal morphological analysis, such as identifying neuron types and large-scale neuron retrieval, all of which require accurate measuring and efficient matching algorithms. Recently, many studies have been conducted to describe neuronal morphologies quantitatively using predefined measurements. However, hand-crafted features are usually inadequate for distinguishing fine-grained differences among massive neurons. In this article, we propose a novel morphology-aware contrastive graph neural network (MACGNN) for unsupervised neuronal morphological representation learning. To improve the retrieval efficiency in large-scale neuronal morphological datasets, we further propose Hash-MACGNN by introducing an improved deep hash algorithm to train the network end-to-end to learn binary hash representations of neurons. We conduct extensive experiments on the largest dataset, NeuroMorpho, which contains more than 100 000 neurons. The experimental results demonstrate the effectiveness and superiority of our MACGNN and Hash-MACGNN for large-scale neuronal morphological analysis.

Mainstream image/video coding standards, exemplified by the state-of-the-art H.266/VVC, AVS3, and AV1, follow the block-based hybrid coding framework. Due to the block-based framework, encoders designed for these standards are easily optimized for peak signal-to-noise ratio (PSNR) but have difficulties optimizing for the metrics more aligned to perceptual quality, e.g., multi-scale structural similarity (MS-SSIM), since these metrics cannot be accurately evaluated at the small block level. We address this problem by leveraging inspiration from the end-to-end image compression built on deep networks, which is easily optimized through network training for any metric as long as the metric is differentiable. We compared the trained models using the same network structure but different metrics and observed that the models allocate rates in different ratios. We then propose a distillation method to obtain the rate allocation rule from end-to-end image compression models with different metrics and to utilize such a rule in the block-based encoders. We implement the proposed method on the VVC reference software—VTM and the AVS3 reference software—HPM, focusing on intraframe coding. Experimental results show that the proposed method on top of VTM achieves more than 10% BD-rate reduction than the anchor when evaluated with MS-SSIM or LPIPS, which leads to concrete perceptual quality improvement.

Event cameras are a kind of bio-inspired imaging sensor, which asynchronously collect sparse event streams with many advantages. In this paper, we focus on building better and faster event-based object detectors. To this end, we first propose a computationally efficient event representation Hyper Histogram, which adequately preserves both the polarity and temporal information of events. Then we devise an Adaptive Event Conversion module, which converts events into Hyper Histograms according to event density via an adaptive queue. Moreover, we introduce a novel event-based augmentation method Shadow Mosaic, which significantly improves the event sample diversity and enhances the generalization ability of detection models. We equip our proposed modules on three representative object detection models: YOLOv5, Deformable-DETR, and RetinaNet. Experimental results on three event-based detection datasets (1Mpx, Gen1, and MVSEC-NIGHTL21) demonstrate that our proposed approach outperforms other state-of-the-art methods by a large margin, while achieving a much faster running speed (< 14 ms and < 4 ms for 50 ms event data on the 1Mpx and Gen1 datasets).

Establishing pixel-level matches between image pairs is vital for a variety of computer vision applications. How-ever, achieving robust image matching remains challenging because CNN extracted descriptors usually lack discrim-inative ability in texture-less regions and keypoint detec-tors are only good at identifying keypoints with a specific level of structure. To deal with these issues, a novel im-age matching method is proposed by Jointly Learning Hier-archical Detectors and Contextual Descriptors via Agent-based Transformers (D2Former), including a contextual feature descriptor learning (CFDL) module and a hierar-chical keypoint detector learning (HKDL) module. The proposed D2Former enjoys several merits. First, the pro-posed CFDL module can model long-range contexts effi-ciently and effectively with the aid of designed descriptor agents. Second, the HKDL module can generate keypoint detectors in a hierarchical way, which is helpful for detecting keypoints with diverse levels of structures. Extensive experimental results on four challenging benchmarks show that our proposed method significantly outperforms state-of-the-art image matching methods.

Camouflaged Instance Segmentation (CIS) aims at predicting the instance-level masks of camouflaged objects, which are usually the animals in the wild adapting their appearance to match the surroundings. Previous instance segmentation methods perform poorly on this task as they are easily disturbed by the deceptive camouflage. To address these challenges, we propose a novel De-camouflaging Network (DCNet) including a pixel-level camouflage decoupling module and an instance-level camouflage suppression module. The proposed DCNet enjoys several merits. First, the pixel-level camouflage decoupling module can extract camouflage characteristics based on the Fourier transformation. Then a difference attention mechanism is proposed to eliminate the camouflage characteristics while reserving target object characteristics in the pixel feature. Second, the instance-level camouflage suppression module can aggregate rich instance information from pixels by use of instance prototypes. To mitigate the effect of background noise during segmentation, we introduce some reliable reference points to build a more robust similarity measurement. With the aid of these two modules, our DCNet can effectively model de-camouflaging and achieve accurate segmentation for camouflaged instances. Extensive experimental results on two benchmarks demonstrate that our DCNet performs favorably against state-of-the-art CIS methods, e.g., with more than 5% performance gains on COD10K and NC4K datasets in average precision.

Automatic mitochondria segmentation enjoys great popularity with the development of deep learning. However, existing methods rely heavily on the labor-intensive manual gathering by experienced domain experts. And naively applying semi-supervised segmentation methods in the natural image field to mitigate the labeling cost is undesirable. In this work, we analyze the gap between mitochondrial images and natural images and rethink how to achieve effective semi-supervised mitochondria segmentation, from the perspective of reliable prototype-level supervision. We propose a novel end-to-end dual-reliable (DualRel) network, including a reliable pixel aggregation module and a reliable prototype selection module. The proposed DualRel enjoys several merits. First, to learn the prototypes well without any explicit supervision, we carefully design the referential correlation to rectify the direct pairwise correlation. Second, the reliable prototype selection module is responsible for further evaluating the reliability of prototypes in constructing prototype-level consistency regularization. Extensive experimental results on three challenging benchmarks demonstrate that our method performs favorably against state-of-the-art semi-supervised segmentation methods. Importantly, with extremely few samples used for training, DualRel is also on par with current state-of-the-art fully supervised methods.

Local feature matching aims at finding correspondences between a pair of images. Although current detector-free methods leverage Transformer architecture to obtain an impressive performance, few works consider maintaining local consistency. Meanwhile, most methods struggle with large scale variations. To deal with the above issues, we propose Adaptive Spot-Guided Transformer (ASTR) for local feature matching, which jointly models the local consistency and scale variations in a unified coarse-to-fine architecture. The proposed ASTR enjoys several merits. First, we design a spot-guided aggregation module to avoid interfering with irrelevant areas during feature aggregation. Second, we design an adaptive scaling module to adjust the size of grids according to the calculated depth information at fine stage. Extensive experimental results on five standard benchmarks demonstrate that our ASTR performs favorably against state-of-the-art methods. Our code will be released on https://astr2023.github.io.

Sparse instance-level supervision has recently been explored to address insufficient annotation in biomedical instance segmentation, which is easier to annotate crowded instances and better preserves instance completeness for 3D volumetric datasets compared to common semi-supervision. In this paper, we propose a sparsely supervised biomedical instance segmentation framework via cross-representation affinity consistency regularization. Specifically, we adopt two individual networks to enforce the perturbation consistency between an explicit affinity map and an implicit affinity map to capture both feature-level instance discrimination and pixel-level instance boundary structure. We then select the highly confident region of each affinity map as the pseudo label to supervise the other one for affinity consistency learning. To obtain the highly confident region, we propose a pseudo-label noise filtering scheme by integrating two entropy-based decision strategies. Extensive experiments on four biomedical datasets with sparse instance annotations show the state-of-the-art performance of our proposed framework. For the first time, we demonstrate the superiority of sparse instance-level supervision on 3D volumetric datasets, compared to common semi-supervision under the same annotation cost. Code is available at https://github.com/liuxy1103/CRAC.

Event cameras are a type of novel neuromorphic sensor that has been gaining increasing attention. Existing event-based backbones mainly rely on image-based designs to extract spatial information within the image transformed from events, overlooking important event properties like time and polarity. To address this issue, we propose a novel Group-based vision Transformer backbone for Event-based vision, called Group Event Transformer (GET), which decouples temporal-polarity information from spatial information throughout the feature extraction process. Specifically, we first propose a new event representation for GET, named Group Token, which groups asynchronous events based on their timestamps and polarities. Then, GET applies the Event Dual Self-Attention block, and Group Token Aggregation module to facilitate effective feature communication and integration in both the spatial and temporal-polarity domains. After that, GET can be integrated with different downstream tasks by connecting it with various heads. We evaluate our method on four event-based classification datasets (Cifar10-DVS, N-MNIST, N-CARS, and DVS128Gesture) and two event-based object detection datasets (1Mpx and Gen1), and the results demonstrate that GET outperforms other state-of-the-art methods. The code is available at https://github.com/Peterande/GET-Group-Event-Transformer.

Memory-based methods in semi-supervised video object segmentation task achieve competitive performance by performing dense matching between query and memory frames. However, most of the existing methods neglect the fact that videos carry rich temporal information yet redundant spatial information. In this case, direct pixel-level global matching will lead to ambiguous correspondences. In this work, we reconcile the inherent tension of spatial and temporal information to retrieve memory frame information along the object trajectory, and propose a novel and coherent Trajectory Memory Retrieval Network (TMRN) to equip with the trajectory information, including a spatial alignment module and a temporal aggregation module. The proposed TMRN enjoys several merits. First, TMRN is empowered to characterize the temporal correspondence which is in line with the nature of video in a data-driven manner. Second, we elegantly customize the spatial alignment module by coupling SVD initialization with agent-level correlation for representative agent construction and rectifying false matches caused by direct pairwise pixel-level correlation, respectively. Extensive experimental results on challenging benchmarks including DAVIS 2017 validation / test and Youtube-VOS 2018/2019 demonstrate that our TMRN, as a general plugin module, achieves consistent improvements over several leading methods.

Tremendous efforts have been made on document image rectification, but how to learn effective representation of such distorted images is still under-explored. In this paper, we present DocMAE, a novel self-supervised framework for document image rectification. Our motivation is to encode the structural cues in document images by leveraging masked autoencoder to benefit the rectification, i.e., the document boundaries, and text lines. Specifically, we first mask random patches of the background-excluded document images and then reconstruct the missing pixels. With such a self-supervised learning approach, the network is encouraged to learn the intrinsic structure of deformed documents by restoring document boundaries and missing text lines. Transfer performance in the downstream rectification task validates the effectiveness of our method. Extensive experiments are conducted to demonstrate the effectiveness of our method.

The capability of humanoid robots to generate facial expressions is crucial for enhancing interactivity and emotional resonance in human-robot interaction. However, humanoid robots vary in mechanics, manufacturing, and ap-pearance. The lack of consistent processing techniques and the complexity of generating facial expressions pose significant challenges in the field. To acquire solutions with high confidence, it is necessary to enable robots to explore the solution space automatically based on performance feedback. To this end, we designed a physical robot with a human-like appearance and developed a general framework for automatic expression generation using the MAP-Elites algorithm. The main advan-tage of our framework is that it does not only generate facial expressions automatically but can also be customized according to user preferences. The experimental results demonstrate that our framework can efficiently generate realistic facial expressions without hard coding or prior knowledge of the robot kinematics. Moreover, it can guide the solution-generation process in accordance with user preferences, which is desirable in many real-world applications.

Neural connectivity reconstruction aims to understand the function of biological reconstruction and promote basic scientific research. The intricate morphology and densely intertwined branches make it an extremely challenging task. Most previous best-performing methods adopt affinity learning or metric learning. Nevertheless, they either neglect to model explicit voxel semantics caused by implicit optimization or are hysteresis to spatial information. Furthermore, the inherent locality of 3D CNNs limits modeling long-range dependencies, leading to sub-optimal results. In this work, we propose a coherent and unified Appearance Prompt Vision Transformer (APViT) to integrate affinity and metric learning to exploit the complementarity by learning long-range spatial dependencies. The proposed APViT enjoys several merits. First, the extension continuity-aware attention module aims at constructing hierarchical attention customized for neuron extensibility and slice continuity to learn instance voxel semantic context from a global perspective and utilize continuity priors to enhance voxel spatial awareness. Second, the appearance prompt modulator is responsible for leveraging voxel-adaptive appearance knowledge conditioned on affinity rich in spatial information to instruct instance voxel semantics, exploiting the potential of affinity learning to complement metric learning. Extensive experimental results on multiple challenging benchmarks demonstrate that our APViT achieves consistent improvements with huge flexibility under the same post-processing strategy.

Reinforcement learning (RL) has emerged as a promising approach for optimizing traffic signal control (TSC) to ensure the efficient operation of transportation networks. However, the traditional trial-and-error technique in RL is usually impractical in real-world applications. Offline RL, which trains models using pre-collected datasets, is a more practical approach. However, this presents challenges such as suboptimal datasets and limited generalization of pre-trained models. To address this, we propose an offline-to-online RL framework for TSC that pre-trains a generalized model and quickly adapts to new traffic scenarios through online refinement. In the offline stage, we augment the pre-collected datasets to cover a diverse set of possible scenarios and use an offline RL method to pretrain a control model. To ensure generalization, we use FRAP-like network as our base model, which is designed to learn the basic logic for signal control. In the online stage, we introduce a discrepancy measure to tackle inconsistencies between offline pre-trained models and online scenarios and prioritize samples based on it. In the experiments, the proposed approach achieves competitive performance and reduces the training time needed for learning in new scenarios, compared to several baselines.

In the past few years, there has been an explosive surge in the use of machine learning (ML) techniques to address combinatorial optimization (CO) problems, especially mixed-integer linear programs (MILPs). Despite the achievements, the limited availability of real-world instances often leads to sub-optimal decisions and biased solver assessments, which motivates a suite of synthetic MILP instance generation techniques. However, existing methods either rely heavily on expert-designed formulations or struggle to capture the rich features of real-world instances. To tackle this problem, we propose G2MILP, *the first* deep generative framework for MILP instances. Specifically, G2MILP represents MILP instances as bipartite graphs, and applies a masked variational autoencoder to iteratively corrupt and replace parts of the original graphs to generate new ones. The appealing feature of G2MILP is that it can learn to generate novel and realistic MILP instances without prior expert-designed formulations, while preserving the structures and computational hardness of real-world datasets, simultaneously. Thus the generated instances can facilitate downstream tasks for enhancing MILP solvers under limited data availability. We design a suite of benchmarks to evaluate the quality of the generated MILP instances. Experiments demonstrate that our method can produce instances that closely resemble real-world datasets in terms of both structures and computational hardness. The deliverables are released at [https://miralab-ustc.github.io/L2O-G2MILP](https://miralab-ustc.github.io/L2O-G2MILP).

The critical challenge of semi-supervised semantic segmentation lies in how to fully exploit a large volume of unlabeled data to improve the model’s generalization performance for robust segmentation. Existing methods tend to employ certain criteria (weighting function) to select pixel-level pseudo labels. However, the trade-off exists between inaccurate yet utilized pseudo-labels, and correct yet discarded pseudo-labels in these methods when handling pseudo-labels without thoughtful consideration of the weighting function, hindering the generalization ability of the model. In this paper, we systematically analyze the trade-off in previous methods when dealing with pseudo-labels. We formally define the trade-off between inaccurate yet utilized pseudo-labels, and correct yet discarded pseudo-labels by explicitly modeling the confidence distribution of correct and inaccurate pseudo-labels, equipped with a unified weighting function. To this end, we propose Distribution-Aware Weighting (DAW) to strive to minimize the negative equivalence impact raised by the trade-off. We find an interesting fact that the optimal solution for the weighting function is a hard step function, with the jump point located at the intersection of the two confidence distributions. Besides, we devise distribution alignment to mitigate the issue of the discrepancy between the prediction distributions of labeled and unlabeled data. Extensive experimental results on multiple benchmarks including mitochondria segmentation demonstrate that DAW performs favorably against state-of-the-art methods.

While deep reinforcement learning (RL) has been demonstrated effective in solving complex control tasks, sample efficiency remains a key challenge due to the large amounts of data required for remarkable performance. Existing research explores the application of representation learning for data-efficient RL, e.g., learning predictive representations by predicting long-term future states. However, many existing methods do not fully exploit the structural information inherent in sequential state signals, which can potentially improve the quality of long-term decision-making but is difficult to discern in the time domain. To tackle this problem, we propose State Sequences Prediction via Fourier Transform (SPF), a novel method that exploits the frequency domain of state sequences to extract the underlying patterns in time series data for learning expressive representations efficiently. Specifically, we theoretically analyze the existence of structural information in state sequences, which is closely related to policy performance and signal regularity, and then propose to predict the Fourier transform of infinite-step future state sequences to extract such information. One of the appealing features of SPF is that it is simple to implement while not requiring storage of infinite-step future states as prediction targets. Experiments demonstrate that the proposed method outperforms several state-of-the-art algorithms in terms of both sample efficiency and performance.

Image deraining is a challenging task since rain streaks have the characteristics of a spatially long structure and have a complex diversity. Existing deep learning-based methods mainly construct the deraining networks by stacking vanilla convolutional layers with local relations, and can only handle a single dataset due to catastrophic forgetting, resulting in a limited performance and insufficient adaptability. To address these issues, we propose a new image deraining framework to effectively explore nonlocal similarity, and to continuously learn on multiple datasets. Specifically, we first design a patchwise hypergraph convolutional module, which aims to better extract the nonlocal properties with higher-order constraints on the data, to construct a new backbone and to improve the deraining performance. Then, to achieve better generalizability and adaptability in real-world scenarios, we propose a biological brain-inspired continual learning algorithm. By imitating the plasticity mechanism of brain synapses during the learning and memory process, our continual learning process allows the network to achieve a subtle stability-plasticity tradeoff. This it can effectively alleviate catastrophic forgetting and enables a single network to handle multiple datasets. Compared with the competitors, our new deraining network with unified parameters attains a state-of-the-art performance on seen synthetic datasets and has a significantly improved generalizability on unseen real rainy images.

Visual tracking aims to estimate object state in a video sequence, which is challenging when facing drastic appearance changes. Most existing trackers conduct tracking with divided parts to handle appearance variations. However, these trackers commonly divide target objects into regular patches by a hand-designed splitting way, which is too coarse to align object parts well. Besides, a fixed part detector is difficult to partition targets with arbitrary categories and deformations. To address the above issues, we propose a novel adaptive part mining tracker (APMT) for robust tracking via a transformer architecture, including an object representation encoder, an adaptive part mining decoder, and an object state estimation decoder. The proposed APMT enjoys several merits. First, in the object representation encoder, object representation is learned by distinguishing target object from background regions. Second, in the adaptive part mining decoder, we introduce multiple part prototypes to adaptively capture target parts through cross-attention mechanisms for arbitrary categories and deformations. Third, in the object state estimation decoder, we propose two novel strategies to effectively handle appearance variations and distractors. Extensive experimental results demonstrate that our APMT achieves promising results with high FPS. Notably, our tracker is ranked the first place in the VOT-STb2022 challenge.

Wasserstein generative adversarial network (WGAN) has attracted great attention due to its solid mathematical background, i.e., to minimize the Wasserstein distance between the generated distribution and the distribution of interest. In WGAN, the Wasserstein distance is quantitatively evaluated by the discriminator, also known as the critic. The vanilla WGAN trained the critic with the simple Lipschitz condition, which was later shown less effective for modeling complex distributions, like the distribution of natural images. We try to improve the WGAN training by introducing pairwise constraint on the critic, oriented to image restoration tasks. In principle, pairwise constraint is to suggest the critic assign a higher rating to the original (real) image than to the restored (generated) image, as long as such a pair of images are available. We show that such pairwise constraint may be implemented by rectifying the gradients in WGAN training, which leads to the proposed rectified Wasserstein generative adversarial network (ReWaGAN). In addition, we build interesting connections between ReWaGAN and the perception-distortion tradeoff. We verify ReWaGAN on two representative image restoration tasks: single image super-resolution (4× and 8×) and compression artifact reduction, where our ReWaGAN not only beats the vanilla WGAN consistently, but also outperforms the state-of-the-art perceptual quality-oriented methods significantly. Our code and models are publicly available at https://github.com/mahaichuan/ReWaGAN.

Weakly supervised object localization (WSOL) aims to predict both object locations and categories with only image-level class labels. However, most existing methods rely on class-specific image regions for localization, resulting in incomplete object localization. To alleviate this problem, we propose a novel end-to-end task-aware framework with a transformer encoder-decoder architecture (TAFormer) to learn class-agnostic foreground maps, including a representation encoder, a localization decoder, and a classification decoder. The proposed TAFormer enjoys several merits. First, the designed three modules can effectively perform class-agnostic localization and classification in a task-aware manner, achieving remarkable performance for both tasks. Second, an optimal transport algorithm is proposed to provide pixel-level pseudo labels to online refine foreground maps. To the best of our knowledge, this is the first work by exploring a task-aware framework with a transformer architecture and an optimal transport algorithm to achieve accurate object localization for WSOL. Extensive experiments with four backbones on two standard benchmarks demonstrate that our TAFormer achieves favorable performance against state-of-the-art methods. Furthermore, we show that the proposed TAFormer provides higher robustness against adversarial attacks and noisy labels.

Semantic matching models-which assume that entities with similar semantics have similar embeddings-have shown great power in knowledge graph embeddings (KGE). Many existing semantic matching models use inner products in embedding spaces to measure the plausibility of triples and quadruples in static and temporal knowledge graphs. However, vectors that have the same inner products with another vector can still be orthogonal to each other, which implies that entities with similar semantics may have dissimilar embeddings. This property of inner products significantly limits the performance of semantic matching models. To address this challenge, we propose a novel regularizer-namely, DUality-induced RegulArizer (DURA)-which effectively encourages the entities with similar semantics to have similar embeddings. The major novelty of DURA is based on the observation that, for an existing semantic matching KGE model (primal), there is often another distance based KGE model (dual) closely associated with it, which can be used as effective constraints for entity embeddings. Experiments demonstrate that DURA consistently and significantly improves the performance of state-of-the-art semantic matching models on both static and temporal knowledge graph benchmarks.

Existing deep learning based de-raining approaches have resorted to the convolutional architectures. However, the intrinsic limitations of convolution, including local receptive fields and independence of input content, hinder the model's ability to capture long-range and complicated rainy artifacts. To overcome these limitations, we propose an effective and efficient transformer-based architecture for the image de-raining. First, we introduce general priors of vision tasks, i.e., locality and hierarchy, into the network architecture so that our model can achieve excellent de-raining performance without costly pre-training. Second, since the geometric appearance of rainy artifacts is complicated and of significant variance in space, it is essential for de-raining models to extract both local and non-local features. Therefore, we design the complementary window-based transformer and spatial transformer to enhance locality while capturing long-range dependencies. Besides, to compensate for the positional blindness of self-attention, we establish a separate representative space for modeling positional relationship, and design a new relative position enhanced multi-head self-attention. In this way, our model enjoys powerful abilities to capture dependencies from both content and position, so as to achieve better image content recovery while removing rainy artifacts. Experiments substantiate that our approach attains more appealing results than state-of-the-art methods quantitatively and qualitatively.

In weakly supervised (WSAL) and unsupervised temporal action localization (UAL), the target is to simultaneously localize temporal boundaries and identify category labels of actions with only video-level category labels (WSAL) or category numbers in a dataset (UAL) during training. Among existing methods, attention based methods have achieved superior performance in both tasks by highlighting action segments with foreground attention weights. However, without the segment-level supervision on the attention weight learning, the quality of the attention weight hinders the performance of these methods. In this paper, we propose a novel Uncertainty Guided Collaborative Training (UGCT) strategy to alleviate this problem, which mainly includes two key designs: (1) The first design is an online pseudo label generation module, in which the RGB and FLOW streams work collaboratively to learn from each other. (2) The second design is an uncertainty aware learning module, which can mitigate the noise in the generated pseudo labels. These two designs work together to promote the model performance effectively and efficiently by exchanging information between RGB and FLOW streams. Extensive experimental results on two benchmark datasets with three attention based methods demonstrate the effectiveness of the proposed method, e.g, more than 7.0% performance gain for mAP@IoU=0.5 on THUMOS14 dataset.

In recent years, memristive crossbar-based neuromorphic computing systems (NCS) have provided a promising solution to the acceleration of neural networks. However, stuck-at faults (SAFs) in the memristor devices significantly degrade the computing accuracy of NCS. Besides, the memristor suffers from the process variations, causing deviation of the actual programming resistance from its target resistance. In this paper, we propose a reliability-driven design framework for a memristive crossbar-based NCS in combination with general and chip-specific design optimizations. First, we design a general reliability-aware training scheme to enhance the robustness of NCS to SAFs and device variations; a dropconnect-inspired approach is developed to alleviate the impact of SAFs; a new weighted error function, including cross-entropy error (CEE), the$l_{2}$-norm of weights, and the sum of squares of first-order derivatives of CEE with respect to weights, is proposed to obtain a smooth error curve, where the effects of variations are suppressed. Second, given the neural network model generated by the reliability-aware training scheme, we exploit chip-specific mapping and re- training to further improve computation accuracy loss incurred by SAFs. Experimental results show that the proposed method can boost the computation accuracy of NCS and improve the NCS robustness. Note to Practitioners—This work is motivated by the manufacturing reliability problem in a memristive crossbar-based NCS. To enhance the robustness of an NCS to SAFs and device variations, this paper presents a reliability-driven design framework with taking account of both general and chip-specific design optimizations. The experimental results have demonstrated that the proposed framework is superior to the prior arts, and can be easily integrated with existing industrial hardware-based fault tolerance solutions for higher accuracy at lower overhead. Memristive crossbar-based computing system gives hope for the anticipated efficient implementation of artificial neuromorphic networks. With the help of the reliability-driven designs, the computation accuracy is restored, and hence we can expect the wide use of memristive crossbar-based computing system in neuromorphic computing applications.

Transformer-based architectures start to emerge in single image super resolution (SISR) and have achieved promising performance. However, most existing vision Transformer-based SISR methods still have two shortcomings: (1) they divide images into the same number of patches with a fixed size, which may not be optimal for restoring patches with different levels of texture richness; and (2) their position encodings treat all input tokens equally and hence, neglect the dependencies among them. This paper presents a HIPA, which stands for a novel Transformer architecture that progressively recovers the high resolution image using a hierarchical patch partition. Specifically, we build a cascaded model that processes an input image in multiple stages, where we start with tokens with small patch sizes and gradually merge them to form the full resolution. Such a hierarchical patch mechanism not only explicitly enables feature aggregation at multiple resolutions but also adaptively learns patch-aware features for different image regions, e.g., using a smaller patch for areas with fine details and a larger patch for textureless regions. Meanwhile, a new attention-based position encoding scheme for Transformer is proposed to let the network focus on which tokens should be paid more attention by assigning different weights to different tokens, which is the first time to our best knowledge. Furthermore, we also propose a multi-receptive field attention module to enlarge the convolution receptive field from different branches. The experimental results on several public datasets demonstrate the superior performance of the proposed HIPA over previous methods quantitatively and qualitatively. We will share our code and models when the paper is accepted.

The Versatile Video Coding (VVC) standard introduces a block partitioning structure known as quadtree plus nested multi-type tree (QTMTT), which allows more flexible block partitioning compared to its predecessors, like High Efficiency Video Coding (HEVC). Meanwhile, the partition search (PS) process, which is to find out the best partitioning structure for optimizing the rate-distortion cost, becomes far more complicated for VVC than for HEVC. Also, the PS process in VVC reference software (VTM) is not friendly to hardware implementation. We propose a partition map prediction method for fast block partitioning in VVC intra-frame encoding. The proposed method may replace PS totally or be combined with PS partially, thereby achieving adjustable acceleration of the VTM intra-frame encoding. Different from the previous methods for fast block partitioning, we propose to represent a QTMTT-based block partitioning structure by a partition map, which consists of a quadtree (QT) depth map, several multi-type tree (MTT) depth maps, and several MTT direction maps. We then propose to predict the optimal partition map from the pixels through a convolutional neural network (CNN). We propose a CNN structure, known as Down-Up-CNN, for the partition map prediction, where the CNN structure emulates the recursive nature of the PS process. Moreover, we design a post-processing algorithm to adjust the network output partition map, so as to obtain a standard-compliant block partitioning structure. The post-processing algorithm may produce a partial partition tree as well; then based on the partial partition tree, the PS process is performed to obtain the full tree. Experimental results show that the proposed method achieves 1.61× to 8.64× encoding acceleration for the VTM-10.0 intra-frame encoder, with the ratio depending on how much PS is performed. Especially, when achieving 3.89× encoding acceleration, the compression efficiency loss is 2.77% in BD-rate, which is a better tradeoff than the previous methods.

Unsupervised deformable image registration benefits from progressive network structures such as Pyramid and Cascade. However, existing progressive networks only consider the single-scale deformation field in each level or stage and ignore the long-term connection across non-adjacent levels or stages. In this paper, we present a novel unsupervised learning approach named Self-Distilled Hierarchical Network (SDHNet). By decomposing the registration procedure into several iterations, SDHNet generates hierarchical deformation fields (HDFs) simultaneously in each iteration and connects different iterations utilizing the learned hidden state. Specifically, hierarchical features are extracted to generate HDFs through several parallel gated recurrent units, and HDFs are then fused adaptively conditioned on themselves as well as contextual features from the input image. Furthermore, different from common unsupervised methods that only apply similarity loss and regularization loss, SDHNet introduces a novel self-deformation distillation scheme. This scheme distills the final deformation field as the teacher guidance, which adds constraints for intermediate deformation fields on deformation-value and deformation-gradient spaces respectively. Experiments on five benchmark datasets, including brain MRI and liver CT, demonstrate the superior performance of SDHNet over state-of-the-art methods with a faster inference speed and a smaller GPU memory. Code is available at https://github.com/Blcony/SDHNet.

Instead of being observed by human, multimedia data are now more and more fed into machines to perform different kinds of semantic analysis. One image may be analyzed multiple times by different machine vision algorithms for different purposes. While machine vision-oriented image compression has been studied, the existing methods are usually driven by a specific machine vision task, and may not be applicable for the other tasks. We address thetask-genericimage compression, in the hope that an image is compressed once but used multiple times for different tasks, all with satisfactory performance. Our study is based on the end-to-end learned image compression. We focus ourselves on the distortion metric, i.e., finding out a task-agnostic metric to estimate the quality of reconstructed images. On the one hand, we study deep feature distance as the metric, which transforms images into a latent space by a pretrained convolutional network—the latent space is believed to be more aligned to semantics—and calculates distance in the latent space. On the other hand, inspired by the saliency mechanism, we study an importance-weighted pixel distance as the metric, where the weights are generated to reflect the importance of the pixels to semantics. Moreover, we combine the two distances into one metric to investigate their complementary nature. An extensive set of experiments are performed to evaluate these metrics. Experimental results show that using the combined metric performs the best, and leads to 20.79%$\sim$42.69% bits saving under the same semantic analysis performance, compared to using signal fidelity metrics. Interestingly, we observe that using the combined metric also improves the visual quality of the reconstructed images.

Many RGB-T trackers attempt to attain robust feature representation by utilizing an adaptive weighting scheme (or attention mechanism). Different from these works, we propose a new dynamic modality-aware filter generation module (named MFGNet) to boost the message communication between visible and thermal data by adaptively adjusting the convolutional kernels for various input images in practical tracking. Given the image pairs as input, we first encode their features with the backbone network. Then, we concatenate these feature maps and generate dynamic modality-aware filters with two independent networks. The visible and thermal filters will be used to conduct a dynamic convolutional operation on their corresponding input feature maps respectively. Inspired by residual connection, both the generated visible and thermal feature maps will be summarized with input feature maps. The augmented feature maps will be fed into the RoI align module to generate instance-level features for subsequent classification. To address issues caused by heavy occlusion, fast motion and out-of-view, we propose to conduct a joint local and global search by exploiting a new direction-aware target driven attention mechanism. The spatial and temporal recurrent neural network is used to capture the direction-aware context for accurate global attention prediction. Extensive experiments on three large-scale RGB-T tracking benchmark datasets validated the effectiveness of our proposed algorithm.

Underexposed images inevitably suffer severe degradation due to light distortion and noise corruption. Motivated by the limited samples of paired datasets, several unsupervised enhancement methods have been developed. However, these techniques heavily rely on pre-defined fixed lightness and noise removal constraints. Correspondingly, they cannot match the image-specific lightness when performing enhancement and can only refine details in a non-perceptual way. In this paper, we propose an Unsupervised Underexposed Image Enhancement Network (U2IENet) with self-illuminated and perceptual guidance. Specifically, to adjust the illumination for matching the image-specific lightness adaptively, we utilize the bright area of the underexposed image as the self-illuminated guidance to constrain the training process and modulate the features. Meanwhile, we introduce the perceptual guidance as a constraint to remove the noise based on illumination distribution, thus refining the details perceptually. Experiments on both underexposed datasets and public low-light datasets demonstrate the superiority of the proposed approach with higher flexibility over state-of- the-art solutions. In addition, our U2IENet also provides a side function that enables users to adjust the lightness via interactive tuning of a single parameter.

Dynamic Vision Sensor (DVS) can asynchronously output the events reflecting apparent motion of objects with microsecond resolution, and shows great application potential in monitoring and other fields. However, the output event stream of existing DVS inevitably contains background activity noise (BA noise) due to dark current and junction leakage current, which will affect the temporal correlation of objects, resulting in deteriorated motion estimation performance. Particularly, the existing filter-based denoising methods cannot be directly applied to suppress the noise in event stream, since there is no spatial correlation. To address this issue, this paper presents a novel progressive framework, in which a Motion Estimation (ME) module and an Event Denoising (ED) module are jointly optimized in a mutually reinforced manner. Specifically, based on the maximum sharpness criterion, ME module divides the input event into several segments by adaptive clustering in a motion compensating warp field, and captures the temporal correlation of event stream according to the clustered motion parameters. Taking temporal correlation as guidance, ED module calculates the confidence that each event belongs to real activity events, and transmits it to ME module to update energy function of motion segmentation for noise suppression. The two steps are iteratively updated until stable motion segmentation results are obtained. Extensive experimental results on both synthetic and real datasets demonstrate the superiority of our proposed approaches against the State-Of-The-Art (SOTA) methods.

Bluetooth is a short-range wireless communication technology widely used by billions of personal computing, IoT, peripheral, and wearable devices. Bluetooth devices exchange commands and data, such as keyboard/mouse inputs, audio, and files, through a secure communication channel that is established through a pairing process. Due to the sensitivity of those commands and data, security mechanisms, such as encryption, authentication, and authorization, have been developed and adopted in the standards. Nevertheless, vulnerabilities continue to be discovered.

Single-step retrosynthesis is the cornerstone of retrosynthesis planning, which is a crucial task for computer-aided drug discovery. The goal of single-step retrosynthesis is to identify the possible reactants that lead to the synthesis of the target product in one reaction. By representing organic molecules as canonical strings, existing sequence-based retrosynthetic methods treat the product-to-reactant retrosynthesis as a sequence-to-sequence translation problem. However, most of them struggle to identify diverse chemical reactions for a desired product due to the deterministic inference, which contradicts the fact that many compounds can be synthesized through various reaction types with different sets of reactants. In this work, we aim to increase reaction diversity and generate various reactants using discrete latent variables. We propose a novel sequence-based approach, namely RetroDVCAE, which incorporates conditional variational autoencoders into single-step retrosynthesis and associates discrete latent variables with the generation process. Specifically, RetroDVCAE uses the Gumbel-Softmax distribution to approximate the categorical distribution over potential reactions and generates multiple sets of reactants with the variational decoder. Experiments demonstrate that RetroDVCAE outperforms state-of-the-art baselines on both benchmark dataset and homemade dataset. Both quantitative and qualitative results show that RetroDVCAE can model the multi-modal distribution over reaction types and produce diverse reactant candidates.

Automatic lip-reading (ALR) aims to recognize words using visual information from the speaker's lip movements. In this work, we introduce a novel type of sensing device, event cameras, for the task of ALR. Event cameras have both technical and application advantages over conventional cameras for the ALR task because they have higher temporal resolution, less redundant visual information, and lower power consumption. To recognize words from the event data, we propose a novel Multi-grained Spatio-Temporal Features Perceived Network (MSTP) to perceive fine-grained spatio-temporal features from microsecond time-resolved event data. Specifically, a multi-branch network architecture is designed, in which different grained spatio-temporal features are learned by operating at different frame rates. The branch operating on the low frame rate can perceive spatial complete but temporal coarse features. While the branch operating on the high frame rate can perceive spatial coarse but temporal refinement features. And a message flow module is devised to integrate the features from different branches, leading to perceiving more discriminative spatio-temporal features. In addition, we present the first event-based lip-reading dataset (DVS-Lip) captured by the event camera. Experimental results demonstrated the superiority of the proposed model compared to the state-of-the-art event-based action recognition models and video-based lip-reading models.

While the majority of FSL models focus on image classification, the extension to action recognition is rather challenging due to the additional temporal dimension in videos. To address this issue, we propose an end-to-end Motion-modulated Temporal Fragment Alignment Network (MT-FAN) by jointly exploring the task-specific motion modulation and the multi-level temporal fragment alignment for Few-Shot Action Recognition (FSAR). The proposed MT-FAN model enjoys several merits. First, we design a motion modulator conditioned on the learned task-specific motion embeddings, which can activate the channels related to the task-shared motion patterns for each frame. Second, a segment attention mechanism is proposed to automatically discover the higher-level segments for multi-level temporal fragment alignment, which encompasses the frame-to-frame, segment-to-segment, and segment-to-frame alignments. To the best of our knowledge, this is the first work to exploit task-specific motion modulation for FSAR. Extensive experimental results on four standard benchmarks demonstrate that the proposed model performs favorably against the state-of-the-art FSAR methods.

Deep graph neural networks (GNNs) have been shown to be expressive for modeling graph-structured data. Nevertheless, the overstacked architecture of deep graph models makes it difficult to deploy and rapidly test on mobile or embedded systems. To compress over-stacked GNNs, knowledge distillation via a teacher-student architecture turns out to be an effective technique, where the key step is to measure the discrepancy between teacher and student networks with predefined distance functions. However, using the same distance for graphs of various structures may be unfit, and the optimal distance formulation is hard to determine. To tackle these problems, we propose a novel Adversarial Knowledge Distillation framework for graph models named GraphAKD, which adversarially trains a discriminator and a generator to adaptively detect and decrease the discrepancy. Specifically, noticing that the well-captured inter-node and inter-class correlations favor the success of deep GNNs, we propose to criticize the inherited knowledge from node-level and class-level views with a trainable discriminator. The discriminator distinguishes between teacher knowledge and what the student inherits, while the student GNN works as a generator and aims to fool the discriminator. Experiments on nodelevel and graph-level classification benchmarks demonstrate that GraphAKD improves the student performance by a large margin. The results imply that GraphAKD can precisely transfer knowledge from a complicated teacher GNN to a compact student GNN.

Generalization across different environments with the same tasks is critical for successful applications of visual reinforcement learning (RL) in real scenarios. However, visual distractions---which are common in real scenes---from high-dimensional observations can be hurtful to the learned representations in visual RL, thus degrading the performance of generalization. To tackle this problem, we propose a novel approach, namely Characteristic Reward Sequence Prediction (CRESP), to extract the task-relevant information by learning reward sequence distributions (RSDs), as the reward signals are task-relevant in RL and invariant to visual distractions. Specifically, to effectively capture the task-relevant information via RSDs, CRESP introduces an auxiliary task---that is, predicting the characteristic functions of RSDs---to learn task-relevant representations, because we can well approximate the high-dimensional distributions by leveraging the corresponding characteristic functions. Experiments demonstrate that CRESP significantly improves the performance of generalization on unseen environments, outperforming several state-of-the-arts on DeepMind Control tasks with different visual distractions.

Vision-and-Language Navigation needs an agent to navigate to a target location by progressively grounding and following the relevant instruction conditioning on its memory and current observation. Existing works utilize the cross-modal transformer to pass the message between visual modality and textual modality. However, they are still limited to mining the fine-grained matching between the underlying components of trajectories and instructions. Inspired by the significant progress achieved by large-scale pre-training methods, in this paper, we propose CSAP, a new method of Cross-modal Semantic Alignment Pre-training for Vision-and-Language Navigation. It is designed to learn the alignment from trajectory-instruction pairs through two novel tasks, including trajectory-conditioned masked fragment modeling and contrastive semantic-alignment modeling. Specifically, the trajectory-conditioned masked fragment modeling encourages the agent to extract useful visual information to reconstruct the masked fragment. The contrastive semantic-alignment modeling is designed to align the visual representation with corresponding phrase embeddings. By showing experimental results on the benchmark dataset, we demonstrate that transformer architecture-based navigation agent pre-trained with our proposed CSAP outperforms existing methods on both SR and SPL scores.

Magnetic resonance imaging (MRI) with high resolution (HR) provides more detailed information for accurate diagnosis and quantitative image analysis. Despite the significant advances, most existing super-resolution (SR) reconstruction network for medical images has two flaws: 1) All of them are designed in a black-box principle, thus lacking sufficient interpretability and further limiting their practical applications. Interpretable neural network models are of significant interest since they enhance the trustworthiness required in clinical practice when dealing with medical images. 2) most existing SR reconstruction approaches only use a single contrast or use a simple multi-contrast fusion mechanism, neglecting the complex relationships between different contrasts that are critical for SR improvement. To deal with these issues, in this paper, a novel Model-Guided interpretable Deep Unfolding Network (MGDUN) for medical image SR reconstruction is proposed. The Model-Guided image SR reconstruction approach solves manually designed objective functions to reconstruct HR MRI. We show how to unfold an iterative MGDUN algorithm into a novel model-guided deep unfolding network by taking the MRI observation matrix and explicit multi-contrast relationship matrix into account during the end-to-end optimization. Extensive experiments on the multi-contrast IXI dataset and BraTs 2019 dataset demonstrate the superiority of our proposed model.

Thanks to the powerful representation capabilities, transformers have made impressive progress in image restoration. However, existing transformers-based methods do not carefully consider the particularities of image restoration. In general, image restoration requires that an ideal approach should be translation-invariant to the degradation, i.e., the undesirable degradation should be removed irrespective of its position within the image. Furthermore, the local relationships also play a vital role, which should be faithfully exploited for recovering clean images. Nevertheless, most transformers either adopt local attention with the fixed local window strategy or global attention, which unfortunately breaks the translation invariance and causes huge loss of local relationships. To address these issues, we propose an elegant stochastic window strategy for transformers. Specifically, we first introduce the window partition with stochastic shift to replace the original fixed window partition for training. Then, we design a new layer expectation propagation algorithm to efficiently approximate the expectation of the induced stochastic transformer for testing. Our stochastic window transformer not only enjoys powerful representation but also maintains the desired property of translation invariance and locality. Experiments validate the stochastic window strategy consistently improves performance on various image restoration tasks (deraining, denoising and deblurring) by significant margins. The code is available at https://github.com/jiexiaou/Stoformer.

Graph convolutional networks (GCNs)—which are effective in modeling graph structures—have been increasingly popular in knowledge graph completion (KGC). GCN-based KGC models first use GCNs to generate expressive entity representations and then use knowledge graph embedding (KGE) models to capture the interactions among entities and relations. However, many GCN-based KGC models fail to outperform state-of-the-art KGE models though introducing additional computational complexity. This phenomenon motivates us to explore the real effect of GCNs in KGC. Therefore, in this paper, we build upon representative GCN-based KGC models and introduce variants to find which factor of GCNs is critical in KGC. Surprisingly, we observe from experiments that the graph structure modeling in GCNs does not have a significant impact on the performance of KGC models, which is in contrast to the common belief. Instead, the transformations for entity representations are responsible for the performance improvements. Based on the observation, we propose a simple yet effective framework named LTE-KGE, which equips existing KGE models with linearly transformed entity embeddings. Experiments demonstrate that LTE-KGE models lead to similar performance improvements with GCN-based KGC methods, while being more computationally efficient. These results suggest that existing GCNs are unnecessary for KGC, and novel GCN-based KGC models should count on more ablation studies to validate their effectiveness. The code of all the experiments is available on GitHub at https://github.com/MIRALab-USTC/GCN4KGC.

Built on deep networks, end-to-end optimized image compression has made impressive progress in the past few years. Previous studies usually adopt a compressive auto-encoder, where the encoder part first converts image into latent features, and then quantizes the features before encoding them into bits. Both the conversion and the quantization incur information loss, resulting in a difficulty to optimally achieve arbitrary compression ratio. We propose iWave++ as a new end-to-end optimized image compression scheme, in which iWave, a trained wavelet-like transform, converts images into coefficients without any information loss. Then the coefficients are optionally quantized and encoded into bits. Different from the previous schemes, iWave++ is versatile: a single model supports both lossless and lossy compression, and also achieves arbitrary compression ratio by simply adjusting the quantization scale. iWave++ also features a carefully designed entropy coding engine to encode the coefficients progressively, and a de-quantization module for lossy compression. Experimental results show that lossy iWave++ achieves state-of-the-art compression efficiency compared with deep network-based methods; on the Kodak dataset, lossy iWave++ leads to 17.34 percent bits saving over BPG; lossless iWave++ achieves comparable or better performance than FLIF. Our code and models are available at https://github.com/mahaichuan/Versatile-Image-Compression.

With the maturity of visual detection techniques, we are more ambitious in describing visual content with open-vocabulary, fine-grained and free-form language, i.e., the task of image captioning. In particular, we are interested in generating longer, richer and more fine-grained sentences and paragraphs as image descriptions. Image captioning can be translated to the task of sequential language prediction given visual content, where the output sequence forms natural language description with plausible grammar. However, existing image captioning methods focus only on language policy while not visual policy, and thus fail to capture visual context that are crucial for compositional reasoning such as object relationships (e.g., "man riding horse") and visual comparisons (e.g., "small(er) cat"). This issue is especially severe when generating longer sequences such as a paragraph. To fill the gap, we propose a Context-Aware Visual Policy network (CAVP) for fine-grained image-to-language generation: image sentence captioning and image paragraph captioning. During captioning, CAVP explicitly considers the previous visual attentions as context, and decides whether the context is used for the current word/sentence generation given the current visual attention. Compared against traditional visual attention mechanism that only fixes a single visual region at each step, CAVP can attend to complex visual compositions over time. The whole image captioning model-CAVP and its subsequent language policy network-can be efficiently optimized end-to-end by using an actor-critic policy gradient method. We have demonstrated the effectiveness of CAVP by state-of-the-art performances on MS-COCO and Stanford captioning datasets, using various metrics and sensible visualizations of qualitative visual context.

The establishment of large video affective content analysis datasets, such as LIRIS-ACCEDE, opens up the possibility of utilizing the massive representation power of deep neural networks (DNNs) to model the complex process of eliciting affective responses from video viewers. However, label noise in these datasets poses a considerable challenge to both the training and evaluation of DNNs. The optimization of DNNs requires stochastic gradient descent (SGD), but label noise in the training set leads to an inaccurate estimate of the gradient, which may cause the model to converge to a nonoptima. In addition, label noise in the test set renders the results of model evaluation untrustworthy. In this article, we propose a multimodal deep quality embedding network (MMDQEN) for affective video content analysis. Specifically, MMDQEN can infer the latent label and label quality from the noisy training samples so that cleaner supervision signals are provided to the DNN-based affective classifier, and a tractable objective for MMDQEN is derived with variational inference and conditional independence assumption. In addition, to avoid model evaluation bias incurred by the annotation noise in the test set, new test sets based on the original LIRIS-ACCEDE database, which we name LIRIS-ACCEDE-RANK, are established where the samples are ranked according to their label uncertainty level, with corresponding evaluation metrics introduced accordingly to further reveal the performance of different models. Experiments conducted on both the LIRIS-ACCEDE and the LIRIS-ACCEDE-RANK datasets demonstrate the effectiveness of the proposed method.

With the rapid development of screen content video applications, screen content coding (SCC) is urgently needed to be used in commercial codecs. However, the extra encoding complexity introduced by the new SCC tools has posed a great challenge for its practical deployment. In this paper, motivated by our observations that there should be a fine-grained mapping between image content and candidate modes, we propose a two-step fast mode decision method to reduce the encoding complexity. First, we propose to use a convolution neural network (CNN) to automatically extract useful features for fine-grained content classification. Second, we build a precise and concise mapping from CUs to candidate modes by simultaneously considering CU content type, CU size, and mode complexity. Note that the spatial correlations between neighboring CUs and current CU are also utilized in candidate modes derivation. In addition to the two-step fast mode decision method, a content-aware early termination algorithm is further proposed to reduce the encoding complexity. Extensive experiments demonstrate that our method achieves better performance compared with state-of-the-art ones, with 50.13% total encoding complexity reduction and only 0.92% BD-rate increase.

Images captured in clinic such as MRI scans are usually in 3D formats with high spatial resolutions. Existing learning-based models for medical image registration consume large GPU memories and long inference time, which is difficult to be deployed in resource-limited diagnosis scenarios. To address this problem, instead of shrinking the model size as in previous works, we turn to reducing the input resolution of existing registration models and boosting their performance through knowledge distillation. Specifically, we propose a cross-resolution distillation (CRD) scheme, which is designed to train low-resolution models under the guidance of corresponding high-resolution models. Nevertheless, due to the resolution gap between features in high/low-resolution models, straightforward distillation is difficult to apply. To overcome this challenge, we first introduce a feature-shifted teacher (FST) to shift and fuse features of high/low-resolution models. Then, we exploit this teacher model to guide the learning of the low-resolution student model with distillation losses on both features and deformation fields. Finally, we only need to use the distilled student model during inference. Experimental results on four 3D medical image datasets demonstrate that the low-resolution models trained through our CRD scheme use fewer than 20% GPU memories and less than 20% inference time while achieving competitive performance compared with corresponding high-resolution models.

It is still a grand challenge to implement a high-accuracy and high-performance stereo matching algorithm on a resource-limited hardware platform in stereo vision systems. This paper proposes a resource-efficient pipelined hardware architecture with four-cycle time-sharing for the semi-global matching (SGM) algorithm with weighted path cost aggregation. To save hardware resources, we also combined image down-sampling and disparity skipping in the SGM algorithm. The presented architecture is synthesized and implemented on a Zynq-7 FPGA board, which results in a throughput of${1280 \times 960/62.5}$fps with 75 disparity levels at the maximum frequency of 216 MHz. To improve the accuracy of the disparity map at close range, we also adapt the presented architecture with two-cycle time-sharing, and the disparity range is increased to 128, which attains the processing of${1280 \times 960/116}$fps at 200 MHz on VCU-118 FPGA board; the throughput reaches 18245 MDE/s. The result shows that the whole architecture only takes 50465 LUTs, 48046 Registers, 125.5 BRAMs with 128 disparity levels, which is much more efficient than the latest reference work.

Dynamic texture refers to the content in video sequences that is characterized by spatial repetition and temporal variation, such as swaying foliage and flowing water. It is a great challenge to compress the dynamic textures efficiently in the current prediction/transform hybrid video coding framework. However, these textures have little information for machine vision, and human visual perception is less sensitive to the textures than to the structures. Thus, we propose a spatiotemporal generative adversarial network (GAN) based dynamic texture synthesis method for surveillance video coding. We detect and remove the dynamic texture content at encoder side, which is irrelevant to machine vision. We generate the dynamic texture content using the proposed GAN at decoder side, so that the reconstructed videos can be observed by human without deteriorating perceptual quality. Specifically, we design a GAN network to synthesize dynamic textures by exploiting the correlation between spatial and temporal neighbors; we present a surveillance video coding scheme with the dynamic texture detection/synthesis method; we build a high-quality dynamic texture dataset, and we collect a dynamic texture testing dataset that goes beyond the existing video coding test datasets by focusing on surveillance scenes. The proposed video coding scheme has been implemented on top of the High Efficiency Video Coding (HEVC) reference software. Experiments have been conducted to evaluate the quantitative and qualitative performance of the proposed coding scheme. Our method achieves 7.4% and 7.6% bit-rate savings in low-delay-B and low-delay-P settings, respectively, at similar visual quality levels in comparison with HEVC.

Single image super-resolution (SISR) using deep convolutional neural networks (CNNs) achieves the state-of-the-art performance. Most existing SISR models mainly focus on pursuing high peak signal-to-noise ratio (PSNR) and neglect textures and details. As a result, the recovered images are often perceptually unpleasant. To address this issue, in this paper, we propose a texture and detail-preserving network (TDPN), which focuses not only on local region feature recovery but also on preserving textures and details. Specifically, the high-resolution image is recovered from its corresponding low-resolution input in two branches. First, a multi-reception field based branch is designed to let the network fully learn local region features by adaptively selecting local region features in different reception fields. Then, a texture and detail-learning branch supervised by the textures and details decomposed from the ground-truth high resolution image is proposed to provide additional textures and details for the super-resolution process to improve the perceptual quality. Finally, we introduce a gradient loss into the SISR field and define a novel hybrid loss to strengthen boundary information recovery and to avoid overly smooth boundary in the final recovered high-resolution image caused by using only the MAE loss. More importantly, the proposed method is model-agnostic, which can be applied to most off-the-shelf SISR networks. The experimental results on public datasets demonstrate the superiority of our TDPN on most state-of-the-art SISR methods in PSNR, SSIM and perceptual quality. We will share our code on https://github.com/tocaiqing/TDPN.

Intensity inhomogeneity and noise are two common issues in images but inevitably lead to significant challenges for image segmentation and is particularly pronounced when the two issues simultaneously appear in one image. As a result, most existing level set models yield poor performance when applied to this images. To this end, this paper proposes a novel hybrid level set model, named adaptive variational level set model (AVLSM) by integrating an adaptive scale bias field correction term and a denoising term into one level set framework, which can simultaneously correct the severe inhomogeneous intensity and denoise in segmentation. Specifically, an adaptive scale bias field correction term is first defined to correct the severe inhomogeneous intensity by adaptively adjusting the scale according to the degree of intensity inhomogeneity while segmentation. More importantly, the proposed adaptive scale truncation function in the term is model-agnostic, which can be applied to most off-the-shelf models and improves their performance for image segmentation with severe intensity inhomogeneity. Then, a denoising energy term is constructed based on the variational model, which can remove not only common additive noise but also multiplicative noise often occurred in medical image during segmentation. Finally, by integrating the two proposed energy terms into a variational level set framework, the AVLSM is proposed. The experimental results on synthetic and real images demonstrate the superiority of AVLSM over most state-of-the-art level set models in terms of accuracy, robustness and running time.

The 360° image that offers a 360-degree scenario of the world is widely used in virtual reality and has drawn increasing attention. In 360° image compression, the spherical image is first transformed into a planar image with a projection such as equirectangular projection (ERP) and then saved with the existing codecs. The ERP images that represent different circles of latitude with the same number of pixels suffer from the unbalance sampling problem, resulting in inefficiency using planar compression methods, especially for the deep neural network (DNN) based codecs. To tackle this problem, we introduce a latitude adaptive coding scheme for DNNs by allocating variant numbers of codes for different regions according to the latitude on the sphere. Specifically, taking both the number of allocated codes for each region and their entropy into consideration, we introduce a flexible regional adaptive rate loss for region-wise rate controlling. Latitude adaptive constraints are then introduced to prevent spending too many codes on the over-sampling regions. Furthermore, we introduce viewport-based distortion loss by calculating the average distortion on a set of viewports. We optimize and test our model on a large 360° dataset containing 19,790 images collected from the Internet. The experiment results demonstrate the superiority of the proposed latitude adaptive coding scheme. On the whole, our model outperforms the existing image compression standards, including JPEG, JPEG2000, HEVC Intra Coding, and VVC Intra Coding, and helps to save around 15% bits compared to the baseline learned image compression model for planar images.

Visible-infrared person re-identification (VI-ReID) is challenging due to the large modality discrepancy between visible and infrared images. Existing methods mainly focus on learning modality-shared representations by embedding images from different modalities into a common feature space, in which some discriminative modality information is discarded. Different from these methods, in this paper, we propose a novel Modality-Specific Memory Network (MSMNet) to complete the missing modality information and aggregate visible and infrared modality features into a unified feature space for the VI-ReID task. The proposed model enjoys several merits. First, it can exploit the missing modality information to alleviate the modality discrepancy when only the single-modality input is provided. To the best of our knowledge, this is the first work to exploit the missing modality information completion and alleviate the modality discrepancy with the memory network. Second, to guide the learning process of the memory network, we design three effective learning strategies, including feature consistency, memory representativeness and structural alignment. By incorporating these learning strategies in a unified model, the memory network can be well learned to propagate identity-related information between modalities and boost the VI-ReID performance. Extensive experimental results on two standard benchmarks (SYSU-MM01 and RegDB) demonstrate that the proposed MSMNet performs favorably against state-of-the-art methods.

Weakly Supervised Object Localization (WSOL) aims to localize objects with only image-level labels, which has better scalability and practicability than fully supervised methods in the actual deployment. However, a common limitation for available techniques based on classification networks is that they only highlight the most discriminative part of the object, not the entire object. To alleviate this problem, we propose a novel end-to-end part discovery model (PDM) to learn multiple discriminative object parts in a unified network for accurate object localization and classification. The proposed PDM enjoys several merits. First, to the best of our knowledge, it is the first work to directly model diverse and robust object parts by exploiting part diversity, compactness, and importance jointly for WSOL. Second, three effective mechanisms including diversity, compactness, and importance learning mechanisms are designed to learn robust object parts. Therefore, our model can exploit complementary spatial information and local details from the learned object parts, which help to produce precise bounding boxes and discriminate different object categories. Extensive experiments on two standard benchmarks demonstrate that our PDM performs favorably against state-of-the-art WSOL approaches.

Weakly supervised object localization (WSOL) aims at localizing objects with only image-level labels, which has better scalability and practicability than fully supervised methods. However, without pixel-level supervision, existing methods tend to generate rough localization maps, which hinders localization performance. To alleviate this problem, we propose an adversarial transformer network (ATNet), which aims to obtain a well-learned localization model with pixel-level pseudo labels. The proposed ATNet enjoys several merits. First, we design an object transformer ( G ) that can generate localization maps and pseudo labels effectively and dynamically, and a part transformer ( D ) to accurately discriminate detailed local differences between localization maps and pseudo labels. Second, we propose to train G and D via an adversarial process, where G can generate more accurate localization maps approaching pseudo labels to fool D . To the best of our knowledge, this is the first work to explore transformers with adversarial training to obtain a well-learned localization model for WSOL. Extensive experiments with four backbones on two standard benchmarks demonstrate that our ATNet achieves favorable performance against state-of-the-art WSOL methods. Besides, our adversarial training can provide higher robustness against adversarial attacks.

Deformation decomposition serves as a good solution for deformable image registration when the deformation is large. Current deformation decomposition methods can be categorized into cascade-based methods and pyramid-based methods. However, cascade-based methods suffer from heavy computational burdens and long inference time due to their structures of repeated subnetworks, while the effectiveness of pyramid-based methods is constrained by their limited numbers of resolution levels. In this paper, to address both the insufficient and inefficient decomposition problems in current deformation decomposition methods, we propose a recursive decomposition network (RDN) to offer a novel solution for deformable image registration. Stage-wise recursion can efficiently decompose a large deformation into different pyramid estimation stages without using repeated subnetworks like in cascade-based methods. Level-wise recursion can sufficiently decompose the deformation inside each resolution level instead of only one-time estimation like in pyramid-based methods. Extensive experiments and ablation studies on two representative datasets validate the effectiveness and efficiency of our proposed RDN.

Connectivity-based brain region parcellation from functional magnetic resonance imaging (fMRI) data is complicated by heterogeneity among aged and diseased subjects, particularly when the data are spatially transformed to a common space. Here, we propose a group-guided functional brain region parcellation model capable of obtaining subregions from a target region with consistent connectivity profiles across multiple subjects, even when the fMRI signals are kept in their native spaces. The model is based on a joint constrained canonical correlation analysis (JC-CCA) method that achieves group-guided parcellation while allowing the data dimension of the parcellated regions for each subject to vary. We performed extensive experiments on synthetic and real data to demonstrate the superiority of the proposed model compared to other classical methods. When applied to fMRI data of subjects with and without Parkinson's disease (PD) to estimate the subregions in the Putamen, significant between-group differences were found in the derived subregions and the connectivity patterns. Superior classification and regression results were obtained, demonstrating its potential in clinical practice.

Emerging deep learning-based methods have enabled great progress in automatic neuron segmentation from Electron Microscopy (EM) volumes. However, the success of existing methods is heavily reliant upon a large number of annotations that are often expensive and time-consuming to collect due to dense distributions and complex structures of neurons. If the required quantity of manual annotations for learning cannot be reached, these methods turn out to be fragile. To address this issue, in this article, we propose a two-stage, semi-supervised learning method for neuron segmentation to fully extract useful information from unlabeled data. First, we devise a proxy task to enable network pre-training by reconstructing original volumes from their perturbed counterparts. This pre-training strategy implicitly extracts meaningful information on neuron structures from unlabeled data to facilitate the next stage of learning. Second, we regularize the supervised learning process with the pixel-level prediction consistencies between unlabeled samples and their perturbed counterparts. This improves the generalizability of the learned model to adapt diverse data distributions in EM volumes, especially when the number of labels is limited. Extensive experiments on representative EM datasets demonstrate the superior performance of our reinforced consistency learning compared to supervised learning, i.e., up to 400% gain on the VOI metric with only a few available labels. This is on par with a model trained on ten times the amount of labeled data in a supervised manner. Code is available at https://github.com/weih527/SSNS-Net.

In this paper, we present LensCast, a novel cross-layer video transmission framework for wireless networks, which seamlessly integrates millimeter wave (mmWave) lens multiple-input multiple-output (MIMO) with robust video transmission. LensCast is designed to exploit the video content diversity at the application layer, together with the spatial path diversity of lens antenna array at the physical layer, to achieve graceful video transmission performance under varying channel conditions. In LensCast, a transmission distortion minimization problem is formulated with the consideration of video chunk scheduling, path matching and power allocation, which is an intractable mixed integer non-linear programming (MINLP) problem. The solution of this MINLP problem is converted into resource allocation (i.e., joint path matching and power allocation) plus chunk scheduling. First, resource allocation is investigated with given chunk scheduling results. By analyzing the optimality of the resource allocation problem, a winner-takes-all assignment is obtained to guide resource allocation. After that, a greedy water-filling algorithm is proposed as a near-optimal solution. Second, we propose a low-complexity chunk scheduling algorithm to schedule chunks for each transmission. Simulation results demonstrate that the proposed LensCast achieves an improved performance in terms of both peak signal-to-noise ratio and visual quality comparing with reference schemes.

The large data volume of point clouds poses severe challenges for efficient storage and transmission in recent years. In this paper, we propose the first--to our best knowledge--end-to-end deep framework for compressing point cloud attributes. Specifically, we propose a point cloud lossy attribute autoencoder, which directly encodes and decodes point cloud attributes with the help of geometry, instead of voxelizing or projecting the points. In the autoencoder, we propose a second-order point convolution that utilizes the spatial correlations between more points and the nonlinear relationship between attribute features. We introduce a dense point-inception block, which derives from a combination of an inception-style block and a dense block, to improve feature propagation. In addition, we devise a multiscale loss to guide the autoencoder in focusing attention on the coarse-grained points with better coverage of the entire point cloud, which makes it easier for the autoencoder to obtain better optimization of the qualities of all points. Experimental results show that our proposed framework still has a performance gap compared with the state-of-the-art algorithms in the MPEG G-PCC reference software TMC13. However, it does outperform the RAHT-RLGR, which is one of the core transforms used in TMC13 without many well-designed techniques that make TMC13 what it is today. It outperforms RAHT-RLGR by 2.63 dB, 1.77 dB, and 3.40 dB on average in terms of the BD-PSNR for the Y, U, and V components. A subjective quality comparison demonstrates that our framework can preserve more textures and reduce blocking and color noise artifacts.

Tracking-by-detection is a very popular framework for single-object tracking that attempts to search the target object within a local search window for each frame. Although such a local search mechanism works well on simple videos, however, it makes the trackers sensitive to extremely challenging scenarios, such as heavy occlusion and fast motion. In this article, we propose a novel and general target-aware attention mechanism (termed TANet) and integrate it with a tracking-by-detection framework to conduct joint local and global search for robust tracking. Specifically, we extract the features of the target object patch and continuous video frames; then, we concatenate and feed them into a decoder network to generate target-aware global attention maps. More importantly, we resort to adversarial training for better attention prediction. The appearance and motion discriminator networks are designed to ensure its consistency in spatial and temporal views. In the tracking procedure, we integrate target-aware attention with multiple trackers by exploring candidate search regions for robust tracking. Extensive experiments on both short- and long-term tracking benchmark datasets all validated the effectiveness of our algorithm.

Brain network is a large-scale complex network with scale-free, small-world, and modularity properties, which largely supports this high-efficiency massive system. In this article, we propose to synthesize brain-network-inspired interconnections for large-scale network-on-chips. First, we propose a method to generate brain-network-inspired topologies with limited scale-free and power-law small-world properties, which have a low total link length and extremely low average hop count approximately proportional to the logarithm of the network size. In addition, given the large-scale applications, considering the modularity of the brain-network-inspired topologies, we present an application mapping method, including task mapping and deterministic deadlock-free routing, to minimize the power consumption and hop count. Finally, a cycle-accurate simulator BookSim2 is used to validate the architecture performance with different synthetic traffic patterns and large-scale test cases, including real-world communication networks for the graph processing application. Experiments show that, compared with other topologies and methods, the brain-network-inspired network-on-chips (NoCs) generated by the proposed method present significantly lower average hop count and lower average latency. Especially in graph processing applications with a power-law and tightly coupled inter-core communication, the brain-network-inspired NoC has up to 70% lower average hop count and 75% lower average latency than mesh-based NoCs.

Inductive link prediction---where entities during training and inference stages can be different---has been shown to be promising for completing continuously evolving knowledge graphs. Existing models of inductive reasoning mainly focus on predicting missing links by learning logical rules. However, many existing approaches do not take into account semantic correlations between relations, which are commonly seen in real-world knowledge graphs. To address this challenge, we propose a novel inductive reasoning approach, namely TACT, which can effectively exploit Topology-Aware CorrelaTions between relations in an entity-independent manner. TACT is inspired by the observation that the semantic correlation between two relations is highly correlated to their topological structure in knowledge graphs. Specifically, we categorize all relation pairs into several topological patterns, and then propose a Relational Correlation Network (RCN) to learn the importance of the different patterns for inductive link prediction. Experiments demonstrate that TACT can effectively model semantic correlations between relations, and significantly outperforms existing state-of-the-art methods on benchmark datasets for the inductive link prediction task.

The fast development of neuromorphic hardwares promotes Spiking Neural Networks (SNNs) to a thrilling research avenue. Current SNNs, though much efficient, are less effective compared with leading Artificial Neural Networks (ANNs) especially in supervised learning tasks. Recent efforts further demonstrate the potential of SNNs in supervised learning by introducing approximated backpropagation (BP) methods. To deal with the non-differentiable spike function in SNNs, these BP methods utilize information from the spatio-temporal domain to adjust the model parameters. With the increasing of time window and network size, the computational complexity of spatio-temporal backpropagation augments dramatically. In this paper, we propose a new backpropagation method for SNNs based on the accumulated spiking flow (ASF), i.e. ASF-BP. In the proposed ASF-BP method, updating parameters does not rely on the spike train of spiking neurons but leverage accumulated inputs and outputs of spiking neurons over the time window, which reduces the BP complexity significantly. We further present an adaptive linear estimation model to approach the dynamic characteristics of spiking neurons statistically. Experimental results demonstrate that with our proposed ASF-BP method, light-weight convolutional SNNs achieve superior performances compared with other spike-based BP methods on both non-neuromorphic (MNIST, CIFAR10) and neuromorphic (CIFAR10-DVS) datasets. The code is available at https://github.com/neural-lab/ASF-BP.

Occluded person re-identification (Re-ID) is a challenging task as persons are frequently occluded by various obstacles or other persons, especially in the crowd scenario. To address these issues, we propose a novel end-to-end Part-Aware Transformer (PAT) for occluded person Re-ID through diverse part discovery via a transformer encoder-decoder architecture, including a pixel context based transformer encoder and a part prototype based transformer decoder. The proposed PAT model enjoys several merits. First, to the best of our knowledge, this is the first work to exploit the transformer encoder-decoder architecture for occluded person Re-ID in a unified deep model. Second, to learn part prototypes well with only identity labels, we design two effective mechanisms including part diversity and part discriminability. Consequently, we can achieve diverse part discovery for occluded person Re-ID in a weakly supervised manner. Extensive experimental results on six challenging benchmarks for three tasks (occluded, partial and holistic Re-ID) demonstrate that our proposed PAT performs favor-ably against stat-of-the-art methods.

Weakly supervised temporal action localization aims to detect and localize actions in untrimmed videos with only video-level labels during training. However, without frame-level annotations, it is challenging to achieve localization completeness and relieve background interference. In this paper, we present an Action Unit Memory Network (AUMN) for weakly supervised temporal action localization, which can mitigate the above two challenges by learning an action unit memory bank. In the proposed AUMN, two attention modules are designed to update the memory bank adaptively and learn action units specific classifiers. Furthermore, three effective mechanisms (diversity, homogeneity and sparsity) are designed to guide the updating of the memory network. To the best of our knowledge, this is the first work to explicitly model the action units with a memory network. Extensive experimental results on two standard benchmarks (THUMOS14 and ActivityNet) demonstrate that our AUMN performs favorably against state-of-the-art methods. Specifically, the average mAP of IoU thresholds from 0.1 to 0.5 on the THUMOS14 dataset is significantly improved from 47.0% to 52.1%.

Diabetic retinopathy (DR) is the leading cause of permanent blindness in the working-age population. And automatic DR diagnosis can assist ophthalmologists to design tailored treatments for patients, including DR grading and lesion discovery. However, most of existing methods treat DR grading and lesion discovery as two independent tasks, which require lesion annotations as a learning guidance and limits the actual deployment. To alleviate this problem, we propose a novel lesion-aware transformer (LAT) for DR grading and lesion discovery jointly in a unified deep model via an encoder-decoder structure including a pixel relation based encoder and a lesion filter based decoder. The proposed LAT enjoys several merits. First, to the best of our knowledge, this is the first work to formulate lesion discovery as a weakly supervised lesion localization problem via a transformer decoder. Second, to learn lesion filters well with only image-level labels, we design two effective mechanisms including lesion region importance and lesion region diversity for identifying diverse lesion regions. Extensive experimental results on three challenging benchmarks including Messidor-1, Messidor-2 and EyePACS demonstrate that the proposed LAT performs favorably against state-of-the-art DR grading and lesion discovery methods.

Tracking by natural language specification is a new rising research topic that aims at locating the target object in the video sequence based on its language description. Compared with traditional bounding box (BBox) based tracking, this setting guides object tracking with high-level semantic information, addresses the ambiguity of BBox, and links local and global search organically together. Those benefits may bring more flexible, robust and accurate tracking performance in practical scenarios. However, existing natural language initialized trackers are developed and compared on benchmark datasets proposed for tracking-by-BBox, which can’t reflect the true power of tracking-by-language. In this work, we propose a new benchmark specifically dedicated to the tracking-by-language, including a large scale dataset, strong and diverse baseline methods. Specifically, we collect 2k video sequences (contains a total of 1,244,340 frames, 663 words) and split 1300/700 for the train/testing respectively. We densely annotate one sentence in English and corresponding bounding boxes of the target object for each video. We also introduce two new challenges into TNL2K for the object tracking task, i.e., adversarial samples and modality switch. A strong baseline method based on an adaptive local-global-search scheme is proposed for future works to compare. We believe this benchmark will greatly boost related researches on natural language guided tracking.

Weakly supervised temporal action detection aims to localize temporal boundaries of actions and identify their categories simultaneously with only video-level category labels during training. Among existing methods, attention based methods have achieved superior performance by separating action and non-action segments. However, without the segment-level ground-truth supervision, the quality of the attention weight hinders the performance of these methods. To alleviate this problem, we propose a novel Uncertainty Guided Collaborative Training (UGCT) strategy, which mainly includes two key designs: (1) The first design is an online pseudo label generation module, in which the RGB and FLOW streams work collaboratively to learn from each other. (2) The second design is an uncertainty aware learning module, which can mitigate the noise in the generated pseudo labels. These two designs work together to promote the model performance effectively and efficiently by imposing pseudo label supervision on attention weight learning. Experimental results on three state-of-the-art attention based methods demonstrate that the proposed training strategy can significantly improve the performance of these methods, e.g., more than 4% for all three methods in terms of mAP@IoU=0.5 on the THUMOS14 dataset.

Rate adaption is one of the decisive factors for the applications of video compression. However, previous deep video compression methods are usually optimized for a single fixed rate-distortion (R-D) tradeoff. While they can achieve multiple bitrates by training multiple independent models, the realized bitrates are limited to several discrete points on the R-D curve and the storage cost increases proportionally to the number of models. In this paper, we propose a variable-rate scheme for deep video compression, which can achieve continuously variable rate by a single model, i.e., it can reach any point on the R-D curve. In our scheme, two deep auto-encoders are used to compress the residual and the motion vector field respectively, which directly generate the final bitstream. The basic rate adaptation can be achieved by using the R-D tradeoff parameter to deeply modulate all the internal feature maps of the auto-encoders. However, other modules in our scheme, notably motion estimation and motion compensation, also affect the final bitrate indirectly. We further use the R-D tradeoff parameter to modulate them via a conditional map, which effectively improves the compression efficiency. We use a multi-rate-distortion loss function together with a step-by-step training strategy to optimize the entire scheme. Our experiments show that the proposed scheme achieves continuously variable rate by a single model with almost the same compression efficiency as multiple fixed-rate models. The additional parameters and computation of our model are negligible when compared with a single fixed-rate model.

Weakly supervised object localization (WSOL) aims to localize objects with only image-level labels, which has better scalability and practicability than fully supervised methods in the actual deployment. However, with only image-level labels, learning object classification models tends to activate object parts and ignore the whole object, while expanding object parts into the whole object may deteriorate classification performance. To alleviate this problem, we propose foreground activation maps (FAM), whose aim is to optimize object localization and classification jointly via an object-aware attention module and a part-aware attention module in a unified model, where the two tasks can complement and enhance each other. To the best of our knowledge, this is the first work that can achieve remarkable performance for both tasks by optimizing them jointly via FAM for WSOL. Besides, the designed two modules can effectively highlight foreground objects for localization and discover discriminative parts for classification. Extensive experiments with four backbones on two standard benchmarks demonstrate that our FAM performs favorably against state-of-the-art WSOL methods.

Few-Shot Learning (FSL) aims at classifying samples into new unseen classes with only a handful of labeled samples available. However, most of the existing methods are based on the image-level pooled representation, yet ignore considerable local clues that are transferable across tasks. To address this issue, we propose an end-to-end Task-aware Part Mining Network (TPMN) by integrating an automatic part mining process into the metric-based model for FSL. The proposed TPMN model enjoys several merits. First, we design a meta filter learner to generate task-aware part filters based on the task embedding in a meta-learning way. The task-aware part filters can adapt to any individual task and automatically mine task-related local parts even for an unseen task. Second, an adaptive importance generator is proposed to identify key local parts and assign adaptive importance weights to different parts. To the best of our knowledge, this is the first work to automatically exploit the task-aware local parts in a meta-learning way for FSL. Extensive experimental results on four standard benchmarks demonstrate that the proposed model performs favorably against state-of-the-art FSL methods.

Rate adaption is one of the decisive factors for the applications of video compression. Previous deep video compression methods are usually optimized for a single fixed rate-distortion (R-D) tradeoff. While they can achieve multiple bitrates by training multiple independent models, the achievable bitrates are limited to several discrete points on the R-D curve and the storage cost increases proportionally to the number of models. We propose a variable-rate scheme for deep video compression, which can achieve continuously variable rate by a single model, i.e., reaching any point on the R-D curve. In our scheme, two deep auto-encoders are used to compress the residual and the motion vector field respectively, which directly generate the final bitstream. The basic rate adaptation can be achieved by using the R-D tradeoff parameter to deeply modulate all the internal feature maps of the auto-encoders. In addition, other modules in our scheme, notably motion estimation and motion compensation, also affect the final bitrate indirectly. We further use the R-D tradeoff parameter to modulate them via a conditional map, thereby effectively improving the compression efficiency. We use a multi-rate-distortion loss function together with a step-by-step training strategy to optimize the entire scheme. The experimental results show the proposed scheme achieves continuously variable rate by a single model with almost the same compression efficiency as multiple fixed-rate models. The additional parameters and computation of our model are negligible when compared with a single fixed-rate model.

Due to the lack of labeled data, it is usually difficult for an unsupervised person re-identification (re-ID) model to learn discriminative features. To address this issue, we propose a global-level and patch-level unsupervised feature learning framework that utilizes both global and local information to obtain more discriminative features. For global-level learning, we design a global similarity-based loss (GSL) to leverage the similarities between whole images. Along with a memory-based non-parametric classifier, the GSL pulls credible samples closer to help train a discriminative model. For patch-level learning, we use a patch generation module to produce different patches. Applying the patch-based discriminative feature learning loss and image-level feature learning loss, the patch branch in the network can learn better representative patch features. Combining the global-level learning with patch-level learning, we obtain a more distinguishable re-ID model. Experimental results obtained on Market-1501 and DukeMTMC-reID datasets validate that our method has great superiority and effectiveness in unsupervised person re-ID.

In the current hybrid coding schemes, mean-squared-error is widely used for the rate-distortion optimization, which leads to high peak signal-to-noise ratio but sub-optimal perceptual quality. Although human perception-related measures, like multi-scale structural similarity (MS-SSIM), have been proposed, plugging them into the hybrid coding schemes may be computationally expensive. Recently, end-to-end optimized image compression has demonstrated the advantage of perceptual quality-oriented optimization by simply changing the training loss function. Inspired by this, we propose to distill the “perceptual” knowledge from end-to-end image compression and use the knowledge to enhance the perceptual quality for Versatile Video Coding (VVC) intra coding. For an input image, we obtain the block-level bit allocation via end-to-end image compression, and use the bit allocation to adjust the quantization parameter of VVC intra coding. Being compatible to the VVC standard, our method achieves on average 9.32% BD-rate reduction on the Kodak image set when evaluated by MS-SSIM, compared to the VVC reference software.

High Efficiency Video Coding (HEVC) achieves significant improvement in compression efficiency by introducing quadtree-based block partition. However, in the HEVC reference software–HM, the optimal partition is found by a recursive rate-distortion optimization (RDO) process, which is computationally expensive and not friendly to hardware implementation. We propose a fast block partitioning algorithm using convolutional neural network (CNN) based depth map prediction for HEVC intra coding. We use the depth map to represent the block partition of a coding tree unit (CTU). Then, we design a CNN to predict the depth map for a CTU, and we construct a large-scale dataset to train the CNN. Through the depth map prediction, we obtain a block partitioning structure for the entire CTU, and then we could directly compress each coding unit, getting rid of the recursive RDO process for partitioning. Experimental results show that our proposed method reduces 65.55% encoding time of HM at the cost of 2.02% Bjøntegaard Delta rate (BD-rate) increase on the common test sequences. For 4K sequences, our method achieves 76.97% time saving with 2.89% BD-rate increase.

Compared with 2D facial expression recognition (FER) and 3D FER, 2D+3D FER can handle the effects of illumination changes and pose variations. The combination of 2D texture and 3D attribute information can further improve the performance. However, most existing approaches still face two challenges: the selection of proper networks for extracting multimodal features, and the significance of local features in salient regions for expression classification. To address these challenges, we propose an efficient feature fusion network with masks (FFNet-M) for 2D+3D FER. Each 3D scan is rep-resented by three types of attribute maps (i.e., depth map, normal map, and texture image), which are then fed into FFNet-M with different networks to extract both 2D and 3D features. Moreover, we design two masks to make FFNet-M focus on 2D local features while paying attention to 3D local features in salient regions. Experimental results show that our FFNet-M outperforms state-of-the-art methods on BU-3DFE dataset and also achieves a high accuracy on Bosphorus dataset.

Deep learning-based dense object detectors have achieved great success in the past few years and have been applied to numerous multimedia applications such as video understanding. However, the current training pipeline for dense detectors is compromised to lots of conjunctions that may not hold. In this paper, we investigate three such important conjunctions: 1) only samples assigned as positive in classification head are used to train the regression head; 2) classification and regression share the same input feature and computational fields defined by the parallel head architecture; and 3) samples distributed in different feature pyramid layers are treated equally when computing the loss. We first carry out a series of pilot experiments to show disentangling such conjunctions can lead to persistent performance improvement. Then, based on these findings, we propose Disentangled Dense Object Detector (DDOD), in which simple and effective disentanglement mechanisms are designed and integrated into the current state-of-the-art dense object detectors. Extensive experiments on MS COCO benchmark show that our approach can lead to 2.0~mAP, 2.4~mAP and 2.2~mAP absolute improvements on RetinaNet, FCOS, and ATSS baselines with negligible extra overhead. Notably, our best model reaches 55.0 mAP on the COCOtest-dev set and 93.5 AP on the hard subset of WIDER FACE, achieving new state-of-the-art performance on these two competitive benchmarks. Code is available at https://github.com/zehuichen123/DDOD.

Albeit existing deep learning-based image de-raining methods have achieved promising results, most of them only extract single scale features, and neglect the fact that similar rain streaks appear repeatedly across different scales. Therefore, this paper aims to explore the cross-scale cues in a multi-scale fashion. Specifically, we first introduce an adaptive-kernel pyramid to provide effective multi-scale information. Then, we design two cross-scale similarity attention blocks (CSSABs) to search spatial and channel relationships between two scales, respectively. The spatial CSSAB explores the spatial similarity between pixels of cross-scale features, while the channel CSSAB emphasizes the interdependencies among cross-scale features. To further improve the diversity of features, we adopt the wavelet transformation and multi-head mechanism in CSSABs to generate multifocal features which focus on different areas. Finally, based on our CSSABs, we construct an effective multifocal attention-based cross-scale network, which exhaustively utilizes the cross-scale correlations of both rain streaks and background, to achieve image de-raining. Experiments show the superiority of our network over state-of-the-art image de-raining approaches both qualitatively and quantitatively. The source code and pre-trained models are available at https://github.com/zhangzheyu0/Multifocal_derain.

Query embedding (QE)---which aims to embed entities and first-order logical (FOL) queries in low-dimensional spaces---has shown great power in multi-hop reasoning over knowledge graphs. Recently, embedding entities and queries with geometric shapes becomes a promising direction, as geometric shapes can naturally represent answer sets of queries and logical relationships among them. However, existing geometry-based models have difficulty in modeling queries with negation, which significantly limits their applicability. To address this challenge, we propose a novel query embedding model, namely \textbf{Con}e \textbf{E}mbeddings (ConE), which is the first geometry-based QE model that can handle all the FOL operations, including conjunction, disjunction, and negation. Specifically, ConE represents entities and queries as Cartesian products of two-dimensional cones, where the intersection and union of cones naturally model the conjunction and disjunction operations. By further noticing that the closure of complement of cones remains cones, we design geometric complement operators in the embedding space for the negation operations. Experiments demonstrate that ConE significantly outperforms existing state-of-the-art methods on benchmark datasets.

Summary In this paper, we present the study of interactional arrangements that support the collaboration of headquarters (HQ), field responders, and a computational planning agent in a time‐critical task setting created by a mixed‐reality game. Interactional arrangements define the extent to which control is distributed between the collaborative parties. We provide 2 field trials, one to study an “on‐the‐loop” arrangement in which HQ monitors and intervenes in agent instructions to field players on demand and the other, to study a version that places HQ more tightly “in‐the‐loop.” The studies provide an understanding of the sociotechnical collaboration between players and the agent in these interactional arrangements by conducting interaction analysis of video recordings and game log data. The first field trial focuses on the collaboration of field responders with the planning agent. Findings highlight how players negotiate the agent guidance within the social interaction of the collocated teams. The second field trial focuses on the collaboration between the automated planning agent and the HQ. We find that the human coordinator and the agent can successfully work together in most cases, with human coordinators inspecting and “correcting” the agent‐proposed plans. Through this field trial‐driven development process, we generalise interaction design implications of automated planning agents around the themes of supporting common ground and mixed‐initiative planning.

A visual scanpath represents the human eye movements when scanning the visual field for acquiring and receiving visual information. Predicting visual scanpaths when a certain stimulus is presented plays an important role in modeling overt human visual attention and search behavior. In this paper, we presented an 'Inhibition of Return - Region of Interest' (IOR-ROI) recurrent mixture density network based framework learning to produce human-like visual scanpaths under task-free viewing conditions. The proposed model simultaneously predicts a sequence of ordered fixation positions and their corresponding fixation durations. Our model integrates bottom-up features and semantic features extracted by convolutional neural networks. Then the integrated feature maps are fed into the IOR-ROI Long Short-Term Memory (LSTM) which is the core component of the proposed model. The IOR-ROI LSTM is a dual LSTM unit, i.e., the IOR-LSTM and the ROI-LSTM, capturing IOR dynamics and gaze shift behavior simultaneously. IOR-LSTM simulates the visual working memory to adaptively maintain and update visual information regarding previously fixated regions. ROI-LSTM is responsible for predicting the next possible ROIs given the spatially inhibited image feature maps on the feature-wise basis. Fixation duration is predicted by a regression neural network given the viewing history and image feature maps corresponding to currently fixated ROI. Considering the eye movement pattern variations among subjects, a mixture density network is adopted to model the next fixation distribution as Gaussian mixtures and the fixation duration is also modeled using Gaussian distribution. Our model is evaluated on the OSIE and MIT low resolution eye-tracking datasets and experimental results indicate that the proposed method can achieve superior performance in predicting visual scanpaths. The code will be publicly available on URL: https://github.com/sunwj/scanpath.

Frame extrapolation is to predict future frames from the past (reference) frames, which has been studied intensively in the computer vision research and has great potential in video coding. Recently, a number of studies have been devoted to the use of deep networks for frame extrapolation, which achieves certain success. However, due to the complex and diverse motion patterns in natural video, it is still difficult to extrapolate frames with high fidelity directly from reference frames. To address this problem, we introduce reference frame alignment as a key technique for deep network-based frame extrapolation. We propose to align the reference frames, e.g. using block-based motion estimation and motion compensation, and then to extrapolate from the aligned frames by a trained deep network. Since the alignment, a preprocessing step, effectively reduces the diversity of network input, we observe that the network is easier to train and the extrapolated frames are of higher quality. We verify the proposed technique in video coding, using the extrapolated frame for inter prediction in High Efficiency Video Coding (HEVC) and Versatile Video Coding (VVC). We investigate different schemes, including whether to align between the target frame and the reference frames, and whether to perform motion estimation on the extrapolated frame. We conduct a comprehensive set of experiments to study the efficiency of the proposed method and to compare different schemes. Experimental results show that our proposal achieves on average 5.3% and 2.8% BD-rate reduction in Y component compared to HEVC, under low-delay P and low-delay B configurations, respectively. Our proposal performs much better than the frame extrapolation without reference frame alignment.

With the rapid development of video distribution and broadcasting, affective video content analysis has attracted a lot of research and development activities recently. Predicting emotional responses of movie audiences is a challenging task in affective computing, since the induced emotions can be considered relatively subjective. In this article, we propose a multimodal local-global attention network (MMLGAN) for affective video content analysis. Inspired by the multimodal integration effect, we extend the attention mechanism to multi-level fusion and design a multimodal fusion unit to obtain a global representation of affective video. The multimodal fusion unit selects key parts from multimodal local streams in the local attention stage and captures the information distribution across time in the global attention stage. Experiments on the LIRIS-ACCEDE dataset, the MediaEval 2015 and 2016 datasets, the FilmStim dataset, the DEAP dataset and the VideoEmotion dataset demonstrate the effectiveness of our approach when compared with the state-of-the-art methods.

End-to-end image compression using trained deep networks as encoding/decoding models has been developed substantially in the recent years. Previous work is limited in using a single encoding/decoding model, whereas we explore the usage of multiple encoding/decoding models as an ensemble. We propose several methods to obtain multiple models. First, we adopt the boosting strategy to train multiple networks with diversity as an ensemble. Second, we train an ensemble of multiple probability distribution models to reduce the distribution gap for efficient entropy coding. Third, we present a geometric transform-based self-ensemble method. The multiple models can be regarded as the multiple coding modes, similar to those in non-deep video coding schemes. We further adopt block-level model/mode selection at the encoder side to pursue rate-distortion optimization, where we use hierarchical block partitioning to improve the adaptation ability. Compared with single-model end-to-end compression, our proposed method improves the compression efficiency significantly, leading to 21% BD-rate reduction on the Kodak dataset, without increasing the decoding complexity. On the other hand, when keeping the same compression efficiency, our method can use much simplified decoding models, where the floating-point operations are reduced by 70%.

Most of the existing single object trackers track the target in a unitary local search window, making them particularly vulnerable to challenging factors such as heavy occlusions and out-of-view movements. Despite the attempts to further incorporate global search, prevailing mechanisms that cooperate local and global search are relatively static, thus are still sub-optimal for improving tracking performance. By further studying the local and global search results, we raise a question: can we allow more dynamics for cooperating both results? In this paper, we propose to introduce more dynamics by devising a dynamic attention-guided multi-trajectory tracking strategy. In particular, we construct dynamic appearance model that contains multiple target templates, each of which provides its own attention for locating the target in the new frame. Guided by different attention, we maintain diversified tracking results for the target to build multi-trajectory tracking history, allowing more candidates to represent the true target trajectory. After spanning the whole sequence, we introduce a multi-trajectory selection network to find the best trajectory that deliver improved tracking performance. Extensive experimental results show that our proposed tracking strategy achieves compelling performance on various large-scale tracking benchmarks. The project page of this paper can be found athttps://sites.google.com/view/mt-track/.

Face restoration from low resolution and noise is important for applications of face analysis recognition. However, most existing face restoration models omit the multiple scale issues in the face restoration problem, which is still not well solved in the research area. In this paper, we propose a sequential gating ensemble network (SGEN) for a multiscale noise robust face restoration issue. To endow the network with multiscale representation ability, we first employ the principle of ensemble learning for SGEN network architecture design. The SGEN aggregates multilevel base-encoders and base-decoders into the network, which enables the network to contain multiple scales of receptive field. Instead of combining these base-en/decoders directly with nonsequential operations, the SGEN takes base-en/decoders from different levels as sequential data. Specifically, it is visualized that SGEN learns to sequentially extract high-level information from base-encoders in a bottom-up manner and restore low-level information from base-decoders in a top-down manner. Besides, we propose realizing bottom-up and top-down information combination and selection with a sequential gating unit (SGU). The SGU sequentially takes information from two different levels as inputs and decides the output based on one active input. Experimental results on the benchmark dataset demonstrate that our SGEN is more effective at multiscale human face restoration with more image details and less noise than state-of-the-art image restoration models. Further utilizing an adversarial training scheme, SGEN also produces more visually preferred results than other models under subjective evaluation.

To establish robust semantic correspondence between images covering different objects belonging to the same category, there are three important types of information including inter-image relationship, intra-image relationship and cycle consistency. Most existing methods only exploit one or two types of the above information and cannot make them enhance and complement each other. Different from existing methods, we propose a novel end-to-end Consistency Graph Modeling Network (CGMNet) for semantic correspondence by modeling inter-image relationship, intra-image relationship and cycle consistency jointly in a unified deep model. The proposed CGMNet enjoys several merits. First, to the best of our knowledge, this is the first work to jointly model the three kinds of information in a deep model for semantic correspondence. Second, our model has designed three effective modules including cross-graph module, intra-graph module and cycle consistency module, which can jointly learn more discriminative feature representations robust to local ambiguities and background clutter for semantic correspondence. Extensive experimental results show that our algorithm performs favorably against state-of-the-art methods on four challenging datasets including PF-PASCAL, PF-WILLOW, Caltech-101 and TSS.

We address the requirement of image coding for joint human-machine vision, i.e., the decoded image serves both human observation and machine analysis/understanding. Previously, human vision and machine vision have been extensively studied by image (signal) compression and (image) feature compression, respectively. Recently, for joint human-machine vision, several studies have been devoted to joint compression of images and features, but the correlation between images and features is still unclear. We identify the deep network as a powerful toolkit for generating structural image representations. From the perspective of information theory, the deep features of an image naturally form an entropy decreasing series: a scalable bitstream is achieved by compressing the features backward from a deeper layer to a shallower layer until culminating with the image signal. Moreover, we can obtain learned representations by training the deep network for a given semantic analysis task or multiple tasks and acquire deep features that are related to semantics. With the learned structural representations, we propose SSSIC, a framework to obtain an embedded bitstream that can be either partially decoded for semantic analysis or fully decoded for human vision. We implement an exemplar SSSIC scheme using coarse-to-fine image classification as the driven semantic analysis task. We also extend the scheme for object detection and instance segmentation tasks. The experimental results demonstrate the effectiveness of the proposed SSSIC framework and establish that the exemplar scheme achieves higher compression efficiency than separate compression of images and features.

Weakly supervised temporal action detection has better scalability and practicability than fully supervised action detection in reality deployment. However, it is difficult to learn a robust model without temporal action boundary annotations. In this paper, we propose an en-to-end Multi-Scale Structure-Aware Network (MSA-Net) for weakly supervised temporal action detection by exploring both the global structure information of a video and the local structure information of actions. The proposed SA-Net enjoys several merits. First, to localize actions with different durations, each video is encoded into feature representations with different temporal scales. Second, based on the multi-scale feature representation, the proposed model has designed two effective structure modeling mechanisms including global structure modeling and local structure modeling, which can effectively learn discriminative structure aware representations for robust and complete action detection. To the best of our knowledge, this is the first work to fully explore the global and local structure information in a unified deep model for weakly supervised action detection. And extensive experimental results on two benchmark datasets demonstrate that the proposed MSA-Net performs favorably against state-of-the-art methods.

Weakly supervised temporal sentence grounding has better scalability and practicability than fully supervised methods in real-world application scenarios. However, most of existing methods cannot model the fine-grained video-text local correspondences well and do not have effective supervision information for correspondence learning, thus yielding unsatisfying performance. To address the above issues, we propose an end-to-end Local Correspondence Network (LCNet) for weakly supervised temporal sentence grounding. The proposed LCNet enjoys several merits. First, we represent video and text features in a hierarchical manner to model the fine-grained video-text correspondences. Second, we design a self-supervised cycle-consistent loss as a learning guidance for video and text matching. To the best of our knowledge, this is the first work to fully explore the fine-grained correspondences between video and text for temporal sentence grounding by using self-supervised learning. Extensive experimental results on two benchmark datasets demonstrate that the proposed LCNet significantly outperforms existing weakly supervised methods.

Recent fMRI connectivity-based parcellation (CBP) methods have been developed to obtain homogeneous and functionally coherent brain parcels. However, most of these studies utilize traditional clustering methods that neglect hidden nonlinear features. To enhance parcellation performance, here we propose a deep embedded connectivity-based parcellation (DECBP) framework and apply it to determine functional subdivisions of the striatum in public resting state fMRI data sets. This framework integrates fMRI connectivity features into deep embedded clustering (DEC), a deep neural network based on a stacked autoencoder. Compared to three prevalent clustering methods and their combinations with principal component analysis (PCA), the DECBP exhibited a significantly higher similarity between scans, individuals, and groups, indicating enhanced reproducibility. The generated reliable parcellations were also largely consistent with other public atlases. We further explored the functional subunits in the striatum in a data set from 23 Parkinson's disease (PD) subjects and 27 age-matched healthy controls (HC). All putaminal subregions of PD demonstrated lower interhemispheric connectivity than those of HC, which might reflect imbalance in the pathological progression of PD. Such hypo-connectivity was also observed between putaminal subregions and other brain regions, reflecting neuroimaging manifestations of the altered cortico-striato-thalamo-cortical circuit. These observed weaker couplings were associated with PD severity and duration. Our results support the utilization of the DECBP framework and suggest that abnormal connectivity in putaminal subregions may be a potential indicator of PD.

Recently, robust video transmission system that can eliminate the cliff effect in digital video transmission has attracted great interest from both academia and industry. By linearizing the whole system, robust video transmission is intrinsically scalable to channel conditions in mobile networks. However, the heterogeneity of user devices in terms of viewing resolution has not been well studied for robust video broadcast systems. In this paper, we propose a spatial scalability enabled robust video broadcast (SSRVB) system, aiming at accommodating diverse users with both heterogeneous resolutions and heterogeneous channel conditions. In SSRVB, a novel spatial decomposition method based on linear projection is first designed for robust video transmission. Then the transmission distortion minimization problem with joint subcarrier matching and power allocation is formulated. A near-optimal low-complexity subcarrier matching algorithm based on auction theory and an optimal power allocation strategy are also proposed. Furthermore, an iterative algorithm is designed to solve the problem of joint resource allocation. Simulation results demonstrate that SSRVB can achieve an average of 3 dB gain when compared with the reference schemes (i.e., ECast, MCast, discrete wavelet transform (DWT) based scheme, and scalable video coding (SVC) scheme) in terms of average peak signal-to-noise ratio under heterogeneous scenarios.

To cope with rapid growth of video services, we propose a non-orthogonal multiple access (NOMA) based scalable video multicast (NOMA-SVM) framework for mobile networks, by exploiting NOMA's specific potential in scalable video multicast transmission. We consider statistical channels, instead of channels with perfect estimation, in the proposed NOMA-SVM framework in order to capture the realistic channel behaviors. As quality of experience (QoE) is a better metric than throughput for video transmission, QoE-driven power allocation is performed among multiple video layers in the proposed NOMA-SVM framework, in which users can decode video with quality proportional to their channel conditions. Specifically, we formulate the power allocation problem with the goal to maximize the average QoE over all users while guaranteeing the basic services of these users. To solve such a non-convex discrete problem, an optimal algorithm is developed based on the hidden monotonicity of the problem. A suboptimal algorithm is also proposed with much lower complexity in order to meet the practical needs. Simulation results show that the proposed algorithms outperform existing orthogonal multiple access (OMA) and NOMA based algorithms under various multicast scenarios in terms of QoE.

Manually labeling neurons from high-resolution but noisy and low-contrast optical microscopy (OM) images is tedious. As a result, the lack of annotated data poses a key challenge when applying deep learning techniques for reconstructing neurons from noisy and low-contrast OM images. While traditional tracing methods provide a possible way to efficiently generate labels for supervised network training, the generated pseudo-labels contain many noisy and incorrect labels, which lead to severe performance degradation. On the other hand, the publicly available dataset, BigNeuron, provides a large number of single 3D neurons that are reconstructed using various imaging paradigms and tracing methods. Though the raw OM images are not fully available for these neurons, they convey essential morphological priors for complex 3D neuron structures. In this paper, we propose a new approach to exploit morphological priors from neurons that have been reconstructed for training a deep neural network to extract neuron signals from OM images. We integrate a deep segmentation network in a generative adversarial network (GAN), expecting the segmentation network to be weakly supervised by pseudo-labels at the pixel level while utilizing the supervision of previously reconstructed neurons at the morphology level. In our morphological-prior-guided neuron reconstruction GAN, named MP-NRGAN, the segmentation network extracts neuron signals from raw images, and the discriminator network encourages the extracted neurons to follow the morphology distribution of reconstructed neurons. Comprehensive experiments on the public VISoR-40 dataset and BigNeuron dataset demonstrate that our proposed MP-NRGAN outperforms state-of-the-art approaches with less training effort.

Multiview learning has shown its superiority in visual classification compared with the single-view-based methods. Especially, due to the powerful representation capacity, the Gaussian process latent variable model (GPLVM)-based multiview approaches have achieved outstanding performances. However, most of them only follow the assumption that the shared latent variables can be generated from or projected to the multiple observations but fail to exploit the harmonization in the back constraint and adaptively learn a classifier according to these learned variables, which would result in performance degradation. To tackle these two issues, in this article, we propose a novel harmonization shared autoencoder GPLVM with a relaxed Hamming distance (HSAGP-RHD). Particularly, an autoencoder structure with the Gaussian process (GP) prior is first constructed to learn the shared latent variable for multiple views. To enforce the agreement among various views in the encoder, a harmonization constraint is embedded into the model by making consistency for the view-specific similarity. Furthermore, we also propose a novel discriminative prior, which is directly imposed on the latent variable to simultaneously learn the fused features and adaptive classifier in a unit model. In detail, the centroid matrix corresponding to the centroids of different categories is first obtained. A relaxed Hamming distance (RHD)-based measurement is subsequently presented to measure the similarity and dissimilarity between the latent variable and centroids, not only allowing us to get the closed-form solutions but also encouraging the points belonging to the same class to be close, while those belonging to different classes to be far. Due to this novel prior, the category of the out-of-sample is also allowed to be simply assigned in the testing phase. Experimental results conducted on three real-world data sets demonstrate the effectiveness of the proposed method compared with state-of-the-art approaches.

Underwater imaging often suffers from color cast and contrast degradation due to range-dependent medium absorption and light scattering. Introducing image statistics as prior has been proved to be an effective solution for underwater image enhancement. However, relative to the modal divergence of light propagation and underwater scenery, the existing methods are limited in representing the inherent statistics of underwater images resulting in color artifacts and haze residuals. To address this problem, this article proposes a convolutional neural network (CNN)-based framework to learn hierarchical statistical features related to color cast and contrast degradation and to leverage them for underwater image enhancement. Specifically, a pixel disruption strategy is first proposed to suppress intrinsic colors’ influence and facilitate modeling a unified statistical representation of underwater image. Then, considering the local variation of depth of field, two parallel sub-networks: Color Correction Network (CC-Net) and Contrast Enhancement Network (CE-Net) are presented. The CC-Net and CE-Net can generate pixel-wise color cast and transmission map and achieve spatial-varied color correction and contrast enhancement. Moreover, to address the issue of insufficient training data, an imaging model-based synthesis method that incorporates pixel disruption strategy is presented to generate underwater patches with global degradation consistency. Quantitative and subjective evaluations demonstrate that our proposed method achieves state-of-the-art performance.

Recently, sentiment analysis has seen remarkable advance with the help of pre-training approaches. However, sentiment knowledge, such as sentiment words and aspect-sentiment pairs, is ignored in the process of pre-training, despite the fact that they are widely used in traditional sentiment analysis approaches. In this paper, we introduce Sentiment Knowledge Enhanced Pre-training (SKEP) in order to learn a unified sentiment representation for multiple sentiment analysis tasks. With the help of automatically-mined knowledge, SKEP conducts sentiment masking and constructs three sentiment knowledge prediction objectives, so as to embed sentiment information at the word, polarity and aspect level into pre-trained sentiment representation. In particular, the prediction of aspect-sentiment pairs is converted into multi-label classification, aiming to capture the dependency between words in a pair. Experiments on three kinds of sentiment tasks show that SKEP significantly outperforms strong pre-training baseline, and achieves new state-of-the-art results on most of the test datasets. We release our code at https://github.com/baidu/Senta.

Camera trace is a unique noise produced in digital imaging process. Most existing forensic methods analyze camera trace to identify image origins. In this paper, we address a new low-level vision problem, camera trace erasing, to reveal the weakness of trace-based forensic methods. A comprehensive investigation on existing anti-forensic methods reveals that it is non-trivial to effectively erase camera trace while avoiding the destruction of content signal. To reconcile these two demands, we propose Siamese Trace Erasing (SiamTE), in which a novel hybrid loss is designed on the basis of Siamese architecture for network training. Specifically, we propose embedded similarity, truncated fidelity, and cross identity to form the hybrid loss. Compared with existing anti-forensic methods, SiamTE has a clear advantage for camera trace erasing, which is demonstrated in three representative tasks.

We propose an end-to-end learned video compression scheme for low-latency scenarios. Previous methods are limited in using the previous one frame as reference. Our method introduces the usage of the previous multiple frames as references. In our scheme, the motion vector (MV) field is calculated between the current frame and the previous one. With multiple reference frames and associated multiple MV fields, our designed network can generate more accurate prediction of the current frame, yielding less residual. Multiple reference frames also help generate MV prediction, which reduces the coding cost of MV field. We use two deep auto-encoders to compress the residual and the MV, respectively. To compensate for the compression error of the auto-encoders, we further design a MV refinement network and a residual refinement network, taking use of the multiple reference frames as well. All the modules in our scheme are jointly optimized through a single rate-distortion loss function. We use a step-by-step training strategy to optimize the entire scheme. Experimental results show that the proposed method outperforms the existing learned video compression methods for low-latency mode. Our method also performs better than H.265 in both PSNR and MS-SSIM. Our code and models are publicly available.

The key of image and sentence matching is to accurately measure the visual-semantic similarity between an image and a sentence. However, most existing methods make use of only the intra-modality relationship within each modality or the inter-modality relationship between image regions and sentence words for the cross-modal matching task. Different from them, in this work, we propose a novel MultiModality Cross Attention (MMCA) Network for image and sentence matching by jointly modeling the intra-modality and inter-modality relationships of image regions and sentence words in a unified deep model. In the proposed MMCA, we design a novel cross-attention mechanism, which is able to exploit not only the intra-modality relationship within each modality, but also the inter-modality relationship between image regions and sentence words to complement and enhance each other for image and sentence matching. Extensive experimental results on two standard benchmarks including Flickr30K and MS-COCO demonstrate that the proposed model performs favorably against state-of-the-art image and sentence matching methods.

Generalized Zero-Shot Learning (GZSL) aims at recognizing both seen and unseen classes by constructing correspondence between visual and semantic embedding. However, existing methods have severely suffered from the strong bias problem, where unseen instances in target domain tend to be recognized as seen classes in source domain. To address this issue, we propose an end-to-end Self-supervised Domain-aware Generative Network (SDGN) by integrating self-supervised learning into feature generating model for unbiased GZSL. The proposed SDGN model enjoys several merits. First, we design a cross-domain feature generating module to synthesize samples with high fidelity based on class embeddings, which involves a novel target domain discriminator to preserve the domain consistency. Second, we propose a self-supervised learning module to investigate inter-domain relationships, where a set of anchors are introduced as a bridge between seen and unseen categories. In the shared space, we pull the distribution of target domain away from source domain, and obtain domain-aware features with high discriminative power for both seen and unseen classes. To our best knowledge, this is the first work to introduce self-supervised learning into GZSL as a learning guidance. Extensive experimental results on five standard benchmarks demonstrate that our model performs favorably against state-of-the-art GZSL methods.

Texture recognition is a challenging visual task since various primitives along with their arrangements can be recognized from a same texture image when perceiving with different contexts. Some recent work building on CNNs exploits orderless aggregating to provide invariance to spatial arrangements. However, these methods ignore the inherent structural property of textures, which is a critical cue for distinguishing and describing texture images in the wild. To address this problem, we propose a novel Deep Structure-Revealed Network (DSR-Net) that leverages spatial dependency among the captured primitives as structural representation for texture recognition. Specifically, a primitive capturing module (PCM) is devised to generate multiple primitives from eight directional spatial contexts, in which deep features are firstly extracted under the constrains of direction map and then encoded based on the similarities of neighborhood. Next, these primitives are associated with a dependence learning module (DLM) to generate structural representation, in which a two-way collaborative relationship strategy is introduced to perceive the spatial dependencies among multiple primitives. At last, the structure-revealed texture representations are integrated with spatial ordered information to achieve real-world texture recognition. Evaluation on the five most challenging texture recognition datasets has demonstrated the superiority of the proposed model against state-of-the-art methods. The structure-revealed performances of DSR-Net are further verified on some extensive experiments, including fine-grained classification and semantic segmentation.

We propose a convolutional neural network-based coefficients prediction (CNNCP) method for intra-predicted residues in the High Efficiency Video Coding (HEVC) standard. In HEVC, discrete cosine transform (DCT) or discrete sine transform (DST) is adopted to convert the intra-predicted residues in the spatial domain into coefficients in the frequency domain. Each coefficient is scalar quantized and entropy coded into the bitstream. As DCT or DST is non-optimal linear transform, there still exist linear and non-linear correlations among different coefficients after the transform. In addition, there exist coefficients' correlations between current block and neighboring blocks, as these correlations cannot be completely exploited in the intra prediction. We thus propose to perform coefficients prediction to further reduce the redundancy among coefficients. The coefficients prediction is achieved using trained convolutional neural networks (CNNs), as CNNs can build complex relationship between input and output by training with a lot of data. In addition, a flag that signals whether to perform coefficients prediction or not at the coding unit level is transmitted to decoder. The proposed CNNCP method is implemented upon the HEVC reference software. Experimental results show that the proposed method achieves on average 1.8%, 4.1%, and 4.5% BD-rate reduction ratios in Y, U, V, respectively, compared with the HEVC baseline in all-intra configuration. In particular, the average BD-rate reduction ratios for 4K test sequences are 2.9%, 6.5%, and 6.6%.

In this paper, we consider a novel image coding paradigm, termed semantically scalable coding. In the new paradigm, coded bitstream serves for multiple different semantic analysis tasks, and different tasks require different semantic granularities of the image. Thus, the bitstream is designed to be scalable in the sense that progressive decoding of the bitstream provides coarse-to-fine semantic granularities. As a concrete example, we consider the task of coarse-grained and fine-grained image classification. We present a method to compress the multiple deep feature maps that are intermediate representations of an image passing a trained deep network. The deep-layer feature maps can serve for coarse-grained image classification while the shallow-layer feature maps can serve for fine-grained image classification. Experimental results demonstrate the feasibility of the proposed method, as well as the advantage of the semantically scalable coding paradigm.

Few-shot classification aims to learn a model that can generalize well to new classes-that are unseen in the training phase-with a small number of labeled instances. Many existing approaches learn a shared embedding function across various tasks to measure the similarities between support (train) and query (test) samples. However, the embeddings generated by these approaches fail to take into account the feature importance of different instances and the feature correlation between support and query samples in each task. To tackle this problem, we propose a novel Self-Adaptive Embedding approach (SAE) by introducing a hierarchical attention scheme. The major novelty of SAE lies in two folds. First, SAE can effectively capture the most discriminative features at the instance level, which significantly improves its performance on downstream classification tasks. Second, SAE can adaptively adjust the representations of support and query samples by considering the feature structures shared by them at the task level. Experiments demonstrate that SAE significantly outperforms existing state-of-the-art methods.

We propose a novel algorithm based on Monte-Carlo tree search for the problem of coalition structure generation (CSG). Specifically, we find the optimal solution by sampling the coalition structure graph and incrementally expanding a search tree, which represents the partial space that has been searched. We prove that our algorithm is complete and converges to the optimal given sufficient number of iterations. Moreover, it is anytime and can scale to large CSG problems with many agents. Experimental results on six common CSG benchmark problems and a disaster response domain confirm the advantages of our approach comparing to the state-of-the-art methods.

In recent years, memristive crossbar-based neuromorphic computing systems (NCS) have provided a promising solution to the acceleration of neural networks. However, stuck-at faults (SAFs) in the memristor devices significantly degrade the computing accuracy of NCS. Besides, the memristor suffers from the process variations, causing deviation of the actual programming resistance from its target resistance. In this paper, we propose a reliability-driven network training framework for a memristive crossbar-based NCS, with taking account of both SAFs and device variations challenges. A dropout-inspired approach is first developed to alleviate the impact of SAFs. A new weighted error function, including cross-entropy error (CEE), the l2-norm of weights, and the sum of squares of first-order derivatives of CEE with respect to weights, is further proposed to obtain a smooth error curve, where the effects of variations are suppressed. Experimental results show that the proposed method can boost the computation accuracy of NCS and improve the NCS robustness.

We propose an image coding scheme that compresses image into semantically scalable bitstream using deep neural networks. This scheme is expected to support intelligent analysis when the bitstream is partially decoded, as well as high-fidelity reconstruction of image when the bitstream is completely decoded. We implement such a semantically scalable image coding scheme based on semantic map. In the proposed scheme, the original image is firstly semantically segmented and the semantic map is compressed as the base layer. Then, the original image is segmented into several individual objects according to the semantic map, and each object is coded separately. A recurrent neural network-based encoder is used to compress these objects at several quality levels. At the decoder side, the semantic map can be directly applied for intelligent analysis. A generative adversarial network is used to synthesize a rough image using the semantic map. If user is interested in a certain object, more bits can be transmitted to enhance the quality of the object. Experimental results show that the proposed method achieves comparable compression performance with JPEG2000 at high bit rates, while facilitates intelligent analysis at low bit rates.

Multi-view learning reveals the latent correlation between different input modalities and has achieved outstanding performances in many fields. Recent approaches aim to find a low-dimensional subspace to reconstruct each view, in which the gross residual or noise follows either Gaussian or Laplacian distribution. However, the noise distribution is often more complex in practical applications, and a deterministic distribution assumption is incapable of modeling it. Additionally, referring to time-changed data, e.g., videos, the noise is temporal smooth, preventing us from processing the data with the whole input, as have generally been done in many existing multi-view learning methods. To tackle these problems, a novel online multi-view subspace learning is proposed in this paper. Particularly, our proposed method not only estimates a transformation for each view to extract the correlation among various views, but also introduces a Mixture of Gausssians (MoG) model into the multi-view data, successfully exploiting numbers of Gaussian Distributions to adaptively fit a wider range of the complex noise. Furthermore, we further design a novel online Expectation Maximization (EM) algorithm, being capable of efficiently processing the dynamic data. Experimental results substantiate the effectiveness and superiority of our approach.

Reward shaping is an effective technique for incorporating domain knowledge into reinforcement learning (RL). Existing approaches such as potential-based reward shaping normally make full use of a given shaping reward function. However, since the transformation of human knowledge into numeric reward values is often imperfect due to reasons such as human cognitive bias, completely utilizing the shaping reward function may fail to improve the performance of RL algorithms. In this paper, we consider the problem of adaptively utilizing a given shaping reward function. We formulate the utilization of shaping rewards as a bi-level optimization problem, where the lower level is to optimize policy using the shaping rewards and the upper level is to optimize a parameterized shaping weight function for true reward maximization. We formally derive the gradient of the expected true reward with respect to the shaping weight function parameters and accordingly propose three learning algorithms based on different assumptions. Experiments in sparse-reward cartpole and MuJoCo environments show that our algorithms can fully exploit beneficial shaping rewards, and meanwhile ignore unbeneficial shaping rewards or even transform them into beneficial ones.

We propose to improve neural network-based compression artifact reduction by transmitting side information for the neural network. The side information consists of artifact descriptors that are obtained by analyzing the original and compressed images in the encoder. In the decoder, the received descriptors are used as additional input to a well-designed conditional post-processing neural network. To reduce the transmission overhead, the entire model is optimized under the rate-distortion constraint via end-to-end learning. Experimental results show that introducing the side information greatly improves the ability of the post-processing neural network, and improves the rate-distortion performance.

Optical flow is a widely used data source for learning motion information, but the complete loss of appearance information limits its ability for action recognition. Therefore we think of enriching optical flow frames with supplementary appearance information to form a new motion data source denoted as Appearance-Supplemented Optical Flow (ASOF). Specifically, we propose a data embedding layer and a stagewise training method to mitigate the scale-wise and density-wise data distribution divergence between the RGB and optical flow frames respectively. We conduct experiments on three benchmark datasets: UCF101 [1], HMDB51 [2] and SomethingSomething-V1 [3]. The results show that our methods can prominently improve optical flow stream recognition accuracy, and further improve the performances of two-stream (score fusion with the RGB stream) with only a little storage increase.

In the hybrid video coding framework, transform is adopted to exploit the dependency within the input signal. In this paper, we propose a deep learning-based nonlinear transform for intra coding. Specifically, we incorporate the directional information into the residual domain. Then, a convolutional neural network model is designed to achieve better decorrelation and energy compaction than the conventional discrete cosine transform. This work has two main contributions. First, we propose to use the intra prediction signal to reduce the directionality in the residual. Second, we present a novel loss function to characterize the efficiency of the transform during the training. To evaluate the compression performance of the proposed transform, we implement it into the High Efficiency Video Coding reference software. Experimental results demonstrate that the proposed method achieves up to 1.79% BD-rate reduction for natural videos.

In video-based point cloud compression (V-PCC), occupancy map video is utilized to indicate whether a 2-D pixel corresponds to a valid 3-D point or not. In the current design of V-PCC, the occupancy map video is directly compressed losslessly with High Efficiency Video Coding (HEVC). However, the coding tools in HEVC are specifically designed for natural images, thus unsuitable for the occupancy map. In this paper, we present a novel quadtree-based scheme for lossless occupancy map coding. In this scheme, the occupancy map is firstly divided into several coding tree units (CTUs). Then, the CTU is divided into coding units (CUs) recursively using a quadtree. The quadtree partition is terminated when one of the three conditions is satisfied. Firstly, all the pixels have the same value. Secondly, the pixels in the CU only have two kinds of values and they can be separated by a continuous edge whose endpoints lie on the side of the CU. The continuous edge is then coded using chain code. Thirdly, the CU reaches the minimum size. This scheme simplifies the design of block partitioning in HEVC and designs simpler yet more effective coding tools. Experimental results show significant reduction of bit-rate and complexity compared with the occupancy map coding scheme in V-PCC. In addition, this scheme is also very efficient to compress the semantic map.

With the advances of Software-Defined Networks (SDN) and Network Function Virtualization (NFV), Service Function Chain (SFC) has been becoming a popular paradigm to carry and complete network services. Such new computing and networking paradigm enables Virtual Network Functions (VNFs) to be placed in software entities/virtual machines over a network of physical equipments in elastic and flexible way with low capital and operation expenses. VNFs are chained together to steer traffic as needed. However, most of the existing traffic steering and routing path computation algorithms for SFC are complex, unscalable, and low time-efficiency. In this paper, we study the VNF Selection and Chaining Problem (VNF-SCP) in SDN/NFV-enabled networks. We formulate VNF-SCP as a Binary Integer Programming (BIP) model in order to compute routing path for each SFC Request (SFCR) with the minimum end-to-end delay. Then, a novel Deep Learning-based Two-Phase Algorithm (DL-TPA) is introduced, where VNF selection network and VNF chaining network are designed to achieve intelligent and efficient VNF selection and chaining for SFCRs. Performance evaluation shows that DL-TPA can achieve high prediction accuracy and time efficiency of routing path computation, and the overall network performance can be improved significantly.

We propose a Deep Boosting Framework (DBF) for real-world image denoising by integrating the deep learning technique into the boosting algorithm. The DBF replaces conventional handcrafted boosting units by elaborate convolutional neural networks, which brings notable advantages in terms of both performance and speed. We design a lightweight Dense Dilated Fusion Network (DDFN) as an embodiment of the boosting unit, which addresses the vanishing of gradients during training due to the cascading of networks while promoting the efficiency of limited parameters. The capabilities of the proposed method are first validated on several representative simulation tasks including non-blind and blind Gaussian denoising and JPEG image deblocking. We then focus on a practical scenario to tackle with the complex and challenging real-world noise. To facilitate leaning-based methods including ours, we build a new Real-world Image Denoising (RID) dataset, which contains 200 pairs of high-resolution images with diverse scene content under various shooting conditions. Moreover, we conduct comprehensive analysis on the domain shift issue for real-world denoising and propose an effective one-shot domain transfer scheme to address this issue. Comprehensive experiments on widely used benchmarks demonstrate that the proposed method significantly surpasses existing methods on the task of real-world image denoising. Code and dataset are available at https://github.com/ngchc/deepBoosting.

The network-on-chips (NoCs)-based communication architecture is a promising candidate for addressing communication bottlenecks in many-core processors and neural network processors. In this article, we consider the generalized fault-tolerance topology generation problem, where the link (physical channel) or switch failures can happen, for application-specific NoCs (ASNoCs). With a user-defined maximum number of faults, K, we propose an integer linear programming (ILP)-based method to generate ASNoC topologies, which can tolerate at most K faults in switches or links. Given the communication requirements between cores and their floorplan, we first propose a convex-cost-flow-based method to solve a core mapping (CM) problem for building connections between the cores and switches. Second, an ILP-based method is proposed to solve the routing path allocation (PA) problem, where K+1 switch-disjoint routing paths are allocated for every communication flow between the cores. Finally, to reduce switch sizes, we propose to share the switch ports for the connections between the cores and switches and formulate the port sharing problem as a clique-partitioning problem, which is solved by iteratively finding a set of the maximum cliques. Additionally, we propose an ILP-based method to simultaneously solve the CM and routing PA problems when only physical link failures are considered. The experimental results show that the power consumption of fault-tolerance topologies increases almost linearly with K because of the routing path redundancy for fault tolerance. When both switch faults and link faults are considered, port sharing can reduce the average power consumption of fault-tolerance topologies with K = 1, K = 2, and K = 3 by 18.08%, 28.88%, and 34.20%, respectively. When considering only the physical link faults, the experimental results show that compared to the fault-tolerant topology generation (FTTG) algorithm, the proposed method reduces power consumption and hop count by 10.58% and 6.25%, respectively; compared to the de Bruijn Digraph (DBG)-based method, the proposed method reduces power consumption and hop count by 21.72% and 9.35%, respectively.

A deep neural network (DNN) based power control method that aims at solving the non-convex optimization problem of maximizing the sum rate of a fading multi-user interference channel is proposed. Towards this end, we first present PCNet, which is a multi-layer fully connected neural network that is specifically designed for the power control problem. A key challenge in training a DNN for the power control problem is the lack of ground truth, i.e., the optimal power allocation is unknown. To address this issue, PCNet leverages the unsupervised learning strategy and directly maximizes the sum rate in the training phase. We then present PCNet+, which enhances the generalization capacity of PCNet by incorporating noise power as an input to the network. Observing that a single PCNet(+) does not universally outperform the existing solutions, we further propose ePCNet(+), a network ensemble with multiple PCNets(+) trained independently. Simulation results show that for the standard symmetric K -user Gaussian interference channel, the proposed methods can outperform state-of-the-art power control solutions under a variety of system configurations. Furthermore, the performance improvement of ePCNet comes with a reduced computational complexity.

One key challenge to learning-based video compression is that motion predictive coding, a very effective tool for video compression, can hardly be trained into a neural network. In this paper, we propose the concept of Pixel-MotionCNN (PMCNN) which includes motion extension and hybrid prediction networks. PMCNN can model spatiotemporal coherence to effectively perform predictive coding inside the learning network. On the basis of PMCNN, we further explore a learning-based framework for video compression with additional components of iterative analysis/synthesis and binarization. The experimental results demonstrate the effectiveness of the proposed scheme. Although entropy coding and complex configurations are not employed in this paper, we still demonstrate superior performance compared with MPEG-2 and achieve comparable results with H.264 codec. The proposed learning-based scheme provides a possible new direction to further improve compression efficiency and functionalities of future video coding.

Recently, pseudo-analog transmission based on SoftCast has been proposed to improve the received quality of video/image by eliminating the cliff effect in traditional digital transmission. In this paper, we propose a Compressed Pseudo-analog Transmission System (ComPaTS) for remote sensing images over bandwidth-constrained wireless channels. This novel scheme is developed based on the observation that the inherent dropping strategy in pseudo-analog transmission is impractical for remote sensing images in which the transmission bandwidth is generally insufficient. In ComPaTS, to guarantee the content diversity gain under pseudo-analog transmission, block-based Compressive Sensing (CS) is applied to the wavelet domain of each remote sensing image, where the sampling ratio is proportional to the importance of different blocks. The main work of ComPaTS is to leverage the sampling ratio in block-based CS and the resource allocation in pseudo-analog transmission in order to minimize system distortion. Two components of system distortion, i.e., source distortion and channel distortion are analyzed respectively. To characterize the coupling relationship between these two different types of distortion, a joint bandwidth-power distortion optimization problem is formulated. Furthermore, we also propose a two-stage allocation algorithm to solve the problem efficiently. The simulation results demonstrate that the proposed ComPaTS scheme significantly outperforms reference schemes in terms of peak signal-to-noise ratio under different bandwidth-constrained scenarios.

Dictionary learning has emerged as a powerful tool for a range of image processing applications and a proper dictionary always plays a key issue to the final achievable performance. In this paper, a class-oriented discriminative dictionary learning (CODDL) method is presented for image classification applications. It takes a comprehensive consideration of multiple optimization objectives, emphasizing class discrimination of both dictionary atoms and representation coefficients. The atoms of the learned dictionary should be grouped into class level sub-dictionaries. Meanwhile, the sparse representation coefficients of an input sample should be concentrated on the sub-dictionary of the class it belongs to. Then, based on the learned class-oriented discriminative dictionary, the structured representation coefficients can thus be used for image classification with a simple and efficient classification scheme. The superior performance of the proposed algorithm is demonstrated through extensive experiments.

Deep learning has achieved great success in the past decade, especially in the fields of computer vision and image processing. After witnessing such success, video coding experts are motivated to consider whether deep learning can also benefit video coding, and if so, they seek to discover why and how. Indeed, a number of research studies have been conducted to explore deep learning for image and video coding, which has been an active and fast-growing research area especially since the year 2015. These prior arts can be divided into two categories: new coding schemes that are built solely upon deep networks (deep schemes), and deep network-based coding tools that are embedded into traditional coding schemes (deep tools). Moreover, in the responses to the joint call for proposals on video compression with capability beyond High Efficiency Video Coding (HEVC), a number of deep tools have been proposed, and some of them are further studied for the upcoming Versatile Video Coding (VVC). In this paper, we summarize the ongoing efforts in the Joint Video Experts Team about the proposed deep tools, and we discuss several promising tools in much detail, including neural network-based intra prediction, convolutional neural network (CNN) based in-loop filtering, and CNN-based block-adaptive-resolution coding. A series of experimental results are provided to demonstrate the capability of these tools in achieving higher compression efficiency than the VVC or HEVC anchor. These results shed light on the promising direction of deep learning-based future video coding, towards which a lot of open problems call for further study.

Entropy coding is a fundamental technology in video coding that removes statistical redundancy among syntax elements. In high efficiency video coding (HEVC), context-adaptive binary arithmetic coding (CABAC) is adopted as the primary entropy coding method. The CABAC consists of three steps: binarization, context modeling, and binary arithmetic coding. As the binarization processes and context models are both manually designed in CABAC, the probability of the syntax elements may not be estimated accurately, which restricts the coding efficiency of CABAC. To address the problem, we propose a convolutional neural network-based arithmetic coding (CNNAC) method and apply it to compress the syntax elements of the intra-predicted residues in HEVC. Instead of manually designing the binarization processes and context models, we propose directly estimating the probability distribution of the syntax elements with a convolutional neural network (CNN), as CNNs can adaptively build complex relationships between inputs and outputs by training with a lot of data. Then, the values of the syntax elements, together with their estimated probability distributions, are fed into a multi-level arithmetic codec to perform entropy coding. In this paper, we have utilized the CNNAC to code the syntax elements of the DC coefficient; the lowest frequency AC coefficient; the second, third, fourth, and fifth lowest frequency AC coefficients; and the position of the last non-zero coefficient in the HEVC intra-predicted residues. The experimental results show that our proposed method achieves up to 6.7% BD-rate reduction and an average of 4.7% BD-rate reduction compared to the HEVC anchor under all intra (AI) configuration.

This paper addresses the challenge of 3D motion recovery by exploiting the spatio-temporal correlations of corrupted 3D skeleton sequences. We propose a new 3D motion recovery method using spatio-temporal reconstruction, which uses joint low-rank and sparse priors to exploit temporal correlation and an isometric constraint for spatial correlation. The proposed model is formulated as a constrained optimization problem, which is efficiently solved by the augmented Lagrangian method with a Gauss-Newton solver for the subproblem of isometric optimization. The experimental results on the CMU motion capture dataset, Edinburgh dataset, and two Kinect datasets demonstrate that the proposed approach achieves better motion recovery than the state-of-the-art methods. The proposed method is applicable to Kinect-like skeleton tracking devices and pose estimation methods that cannot provide accurate estimation of complex motions, especially in the presence of occlusion.

From the non-local self-similarity (NSS)-based image denoising to the convolutional-network (ConvNet)-based image denoising, the denoising performance has been greatly improved. However, it is still not clear how to utilize similar web images to guide image denoising using ConvNet. This paper proposes a novel ConvNet for image denoising to explore both internal (NSS) and external correlations when external similar images are available. Since external similar images may be taken with different viewpoints, focal lengths, and may contain different objects, it is difficult to directly explore external correlations at image level using ConvNet. Therefore, we propose an internal and external patch matching ConvNet (IENet), whose inputs are similar patch cubes extracted from the noisy input and its external similar images. We design three different network structures, namely early-fusion, middle-fusion, and late-fusion of the internal and external cubes to fully combine the strengths of internal and external correlations. The experimental results demonstrate that the proposed method achieves the best denoising results compared with the seven state-of-the-art denoising methods. In specific, the proposed method outperforms the state-of-the-art web image guided denoising method by more than 1 dB on average, which further demonstrates the superiority of the proposed IENet-based filtering over the hand-crafted filtering methods.

Unsupervised person re-identification (Re-ID) has better scalability and practicability than supervised Re-ID in the actual deployment. However, it is difficult to learn a discriminative Re-ID model without annotations. To address the above issue, we propose an end-to-end Self-supervised Agent Learning (SAL) algorithm by exploiting a set of agents as a bridge to reduce domain gaps for unsupervised cross-domain person Re- ID. The proposed SAL model enjoys several merits. First, to the best of our knowledge, this is the first work to exploit selfsupervised learning for unsupervised person Re-ID. Second, our model has designed three effective learning mechanisms including supervised label learning in source domain, similarity consistency learning in target domain, and self-supervised learning in cross domain, which can learn domain-invariant yet discriminative representations through the principled lens of agent learning by reducing domain discrepancy adaptively. Extensive experimental results on three standard benchmarks demonstrate that the proposed SAL performs favorably against state-of-the-art unsupervised person Re-ID methods.

In this paper, a novel deep network is proposed for multi-focus image fusion, named Deep Regression Pair Learning (DRPL). In contrast to existing deep fusion methods which divide the input image into small patches and apply a classifier to judge whether the patch is in focus or not, DRPL directly converts the whole image into a binary mask without any patch operation, subsequently tackling the difficulty of the blur level estimation around the focused/defocused boundary. Simultaneously, a pair learning strategy, which takes a pair of complementary source images as inputs and generates two corresponding binary masks, is introduced into the model, greatly imposing the complementary constraint on each pair and making a large contribution to the performance improvement. Furthermore, as the edge or gradient does exist in the focus part while there is no similar property for the defocus part, we also embed a gradient loss to ensure the generated image to be all-in-focus. Then the structural similarity index (SSIM) is utilized to make a trade-off between the reference and fused images. Experimental results conducted on the synthetic and real-world datasets substantiate the effectiveness and superiority of DRPL compared with other state-of-the-art approaches. The testing code can be found in https://github.com/sasky1/DPRL.

Block transform coded images usually suffer from annoying artifacts at low bit-rates, because of the independent quantization of DCT coefficients. Image prior models play an important role in compressed image reconstruction. Natural image patches in a small neighborhood of the high-dimensional image space usually exhibit an underlying sub-manifold structure. To model the distribution of signal, we extract sub-manifold structure as prior knowledge. We utilize graph Laplacian regularization to characterize the sub-manifold structure at patch level. And similar patches are exploited as samples to estimate distribution of a particular patch. Instead of using Euclidean distance as similarity metric, we propose to use graph-domain distance to measure the patch similarity. Then we perform low-rank regularization on the similar-patch group, and incorporate a non-convex lp penalty to surrogate matrix rank. Finally, an alternatively minimizing strategy is employed to solve the non-convex problem. Experimental results show that our proposed method is capable of achieving more accurate reconstruction than the state-of-the-art methods in both objective and perceptual qualities.

Visually unpleasant compression artifacts frequently appear in block-based transform coding, especially at low bit rates. This paper presents a new artifact reduction scheme based on Bayesian sparse modeling and artifacts-free PCA basis learning. To avoid the effect of blocking artifacts, we propose to learn artifacts-free PCA basis from clean images. We concatenate the clean patches and their compressed counterparts to learn paired distribution prior via the Gaussian Mixture Model (GMM). By this way, the GMM characterizes the mapping between the clean image and its compressed version. To restore a compressed patch, the best matched GMM component is assigned using the patch in the compressed image subspace. The artifacts-free PCA basis is obtained according to the mapping learned by the paired GMM. In practice, the statistical distributions of different sparse coefficients in different patches may dramatically vary with image contents. Instead of using a global zero-mean distribution for all coefficients, we propose to adaptively model the prior of each band in a Bayesian framework. The expectation and variance of each band are adaptively learned from the similar patches within the image. Thus, different transform bands are regularized unequally according to the learned priors. Experimental results show that the proposed scheme outperforms most of the compared schemes in terms of both objective quality and perceptual quality.

Reconstruction of neuronal populations from ultra-scale optical microscopy (OM) images is essential to investigate neuronal circuits and brain mechanisms. The noises, low contrast, huge memory requirement, and high computational cost pose significant challenges in the neuronal population reconstruction. Recently, many studies have been conducted to extract neuron signals using deep neural networks (DNNs). However, training such DNNs usually relies on a huge amount of voxel-wise annotations in OM images, which are expensive in terms of both finance and labor. In this paper, we propose a novel framework for dense neuronal population reconstruction from ultra-scale images. To solve the problem of high cost in obtaining manual annotations for training DNNs, we propose a progressive learning scheme for neuronal population reconstruction (PLNPR) which does not require any manual annotations. Our PLNPR scheme consists of a traditional neuron tracing module and a deep segmentation network that mutually complement and progressively promote each other. To reconstruct dense neuronal populations from a terabyte-sized ultra-scale image, we introduce an automatic framework which adaptively traces neurons block by block and fuses fragmented neurites in overlapped regions continuously and smoothly. We build a dataset "VISoR-40" which consists of 40 large-scale OM image blocks from cortical regions of a mouse. Extensive experimental results on our VISoR-40 dataset and the public BigNeuron dataset demonstrate the effectiveness and superiority of our method on neuronal population reconstruction and single neuron reconstruction. Furthermore, we successfully apply our method to reconstruct dense neuronal populations from an ultra-scale mouse brain slice. The proposed adaptive block propagation and fusion strategies greatly improve the completeness of neurites in dense neuronal population reconstruction.

This article addresses neural network based post-processing for the state-of-the-art video coding standard, High Efficiency Video Coding (HEVC). We first propose a partition-aware convolution neural network (CNN) that utilizes the partition information produced by the encoder to assist in the post-processing. In contrast to existing CNN-based approaches, which only take the decoded frame as input, the proposed approach considers the coding unit (CU) size information and combines it with the distorted decoded frame such that the artifacts introduced by HEVC are efficiently reduced. We further introduce an adaptive-switching neural network (ASN) that consists of multiple independent CNNs to adaptively handle the variations in content and distortion within compressed-video frames, providing further reduction in visual artifacts. Additionally, an iterative training procedure is proposed to train these independent CNNs attentively on different local patch-wise classes. Experiments on benchmark sequences demonstrate the effectiveness of our partition-aware and adaptive-switching neural networks.

Wavelet transform is a powerful tool for multiresolution time-frequency analysis. It has been widely adopted in many image processing tasks, such as denoising, enhancement, fusion, and especially compression. Wavelets lead to the successful image coding standard JPEG-2000. Traditionally, wavelets were designed from the signal processing theory with certain assumption on the signal, but natural images are not as ideal as assumed by the theory. How to design content-adaptive wavelets for natural images remains a difficulty. Inspired by the recent progress of convolutional neural network (CNN), we propose iWave as a framework for deriving wavelet-like transform that is more suitable for natural image compression. iWave adopts an update-first lifting scheme, where the prediction filter is a trained CNN, to achieve wavelet-like transform. The CNN can be embedded into a deep network that is analogous to an auto-encoder, which is trained end-to-end. The trained wavelet-like transform still possesses the lifting structure, which ensures perfect reconstruction, supports multiresolution analysis, and is more interpretable than the deep networks trained as “black boxes.” We perform experiments to verify the generality as well as the speciality of iWave in comparison with JPEG-2000. When trained with a generic set of natural images and tested on the Kodak dataset, iWave achieves on average 4.4% and up to 14% BD-rate reductions. When trained and tested with a specific kind of textures, iWave provides as high as 27% BD-rate reduction.

The newly emerging task of person search with natural language query aims at retrieving the target pedestrian by a text description of the pedestrian. It is more applicable compared to person search with image/video query, i.e., person re-identification. In this paper, we propose a novel Adversarial Attribute-Text Embedding (AATE) network for person search with text query. In particular, a cross-modal adversarial learning module is proposed to learn discriminative and modality-invariant visual-textual features. It consists of a cross-modal learner and a modality discriminator, playing a min-max game in an adversarial learning way. The former is to improve intra-modality discrimination and inter-modality invariance towards confusing the modality discriminator. The latter is to distinguish the features from different modalities and boost the learning of modality-invariant features. Moreover, a visual attribute graph convolutional network is proposed to learn visual attributes of pedestrians, which possess better descriptiveness, interpretability and robustness compared to pedestrian appearance features. A hierarchical text embedding network, consisting of multi-stacked bidirectional LSTMs and a textual attention block, is developed to extract effective textual features from text descriptions of pedestrians. Extensive experimental results on two challenging benchmarks, have demonstrated the effectiveness of the proposed approach.

Due to the powerful capability of the data representation, deep learning has achieved a remarkable performance in supervised hash function learning. However, most of the existing hashing methods focus on point-to-point matching that is too strict and unnecessary. In this article, we propose a novel deep supervised hashing method by relaxing the matching between each pair of instances to a point-to-angle way. Specifically, an inner product is introduced to asymmetrically measure the similarity and dissimilarity between the real-valued output and the binary code. Different from existing methods that strictly enforce each element in the real-valued output to be either +1 or -1, we only encourage the output to be close to its corresponding semantic-related binary code under the cross-angle. This asymmetric product not only projects both the real-valued output and the binary code into the same Hamming space but also relaxes the output with wider choices. To further exploit the semantic affinity, we propose a novel Hamming-distance-based triplet loss, efficiently making a ranking for the positive and negative pairs. An algorithm is then designed to alternatively achieve optimal deep features and binary codes. Experiments on four real-world data sets demonstrate the effectiveness and superiority of our approach to the state of the art.

In recent years, neuromorphic computing systems based on memristive crossbar have provided a promising solution to enable acceleration of neural networks. However, most of the neural networks used in realistic applications are often sparse. If such sparse neural network is directly implemented on a single memristive crossbar, then it would result in inefficient hardware realizations. In this work, we propose E3D-FNC, an enhanced three-dimesnional (3D) floorplanning framework for neuromorphic computing systems, in which the neuron clustering and the layer assignment are considered interactively. First, in each iteration, hierarchical clustering partitions neurons into a set of clusters under the guidance of the proposed distance metric. The optimal number of clusters is determined by L-method. Then matrix re-ordering is proposed to re-arrange the columns of the weight matrix in each cluster. As a result, the reordered connection matrix can be easily mapped into a set of crossbars with high utilizations. Next, since the clustering results will in turn affect the floorplan, we perform the floorplanning of neurons and crossbars again. All the proposed methodologies are embedded in an iterative framework to improve the quality of NCS design. Finally, a 3D floorplan of neuromorphic computing systems is generated. Experimental results show that E3D-FNC can achieve highly hardware-efficient designs compared to the state of the art.

Learning-based methods suffer from a deficiency of clean annotations, especially in biomedical segmentation. Although many semi-supervised methods have been proposed to provide extra training data, automatically generated labels are usually too noisy to retrain models effectively. In this paper, we propose a Two-Stream Mutual Attention Network (TSMAN) that weakens the influence of back-propagated gradients caused by incorrect labels, thereby rendering the network robust to unclean data. The proposed TSMAN consists of two sub-networks that are connected by three types of attention models in different layers. The target of each attention model is to indicate potentially incorrect gradients in a certain layer for both sub-networks by analyzing their inferred features using the same input. In order to achieve this purpose, the attention models are designed based on the propagation analysis of noisy gradients at different layers. This allows the attention models to effectively discover incorrect labels and weaken their influence during parameter updating process. By exchanging multi-level features within two-stream architecture, the effects of noisy labels in each sub-network are reduced by decreasing the noisy gradients. Furthermore, a hierarchical distillation is developed to provide reliable pseudo labels for unlabelded data, which further boosts the performance of TSMAN. The experiments using both HVSMR 2016 and BRATS 2015 benchmarks demonstrate that our semi-supervised learning framework surpasses the state-of-the-art fully-supervised results.

Existing methods for single image super-resolution (SR) are typically evaluated with synthetic degradation models such as bicubic or Gaussian downsampling. In this paper, we investigate SR from the perspective of camera lenses, named as CameraSR, which aims to alleviate the intrinsic tradeoff between resolution (R) and field-of-view (V) in realistic imaging systems. Specifically, we view the R-V degradation as a latent model in the SR process and learn to reverse it with realistic low- and high-resolution image pairs. To obtain the paired images, we propose two novel data acquisition strategies for two representative imaging systems (i.e., DSLR and smartphone cameras), respectively. Based on the obtained City100 dataset, we quantitatively analyze the performance of commonly-used synthetic degradation models, and demonstrate the superiority of CameraSR as a practical solution to boost the performance of existing SR methods. Moreover, CameraSR can be readily generalized to different content and devices, which serves as an advanced digital zoom tool in realistic imaging systems.

The goal of multi-domain learning is to improve the performance over multiple domains by making full use of all training data from them. However, variations of feature distributions across different domains result in a non-trivial solution of multi-domain learning. The state-of-the-art work regarding multi-domain classification aims to extract domain-invariant features and domain-specific features independently. However, they view the distributions of features from different classes as a general distribution and try to match these distributions across domains, which lead to the mixture of features from different classes across domains and degrade the performance of classification. Additionally, existing works only force the shared features among domains to be orthogonal to the features in the domain-specific network. However, redundant features between the domain-specific networks still remain, which may shrink the discriminative ability of domain-specific features. Therefore, we propose an end-to-end network to obtain the more optimal features, which we call compact features. We propose to extract the domain-invariant features by matching the joint distributions of different domains, which have dis- tinct boundaries between different classes. Moreover, we add an orthogonal constraint between the private features across domains to ensure the discriminative ability of the domain-specific space. The proposed method is validated on three landmark datasets, and the results demonstrate the effectiveness of our method.

People who observe a multi-agent team can often provide valuable information to the agents based on their superior cognitive abilities to interpret sequences of observations and assess the overall situation. The knowledge they possess is often difficult to be fully represent using a formal model such as DEC-POMDP. To deal with this, we propose an extension of the DEC-POMDP that allows states to be partially specified and benefit from expert knowledge, while preserving the partial observability and decentralized operation of the agents. In particular, we present an algorithm for computing policies based on history samples that include human labeled data in the form of reward reshaping. We also consider ways to minimize the burden on human experts during the labeling phase. The results offer the first approach to incorporating human knowledge in such complex multi-agent settings. We demonstrate the benefits of our approach using a disaster recovery scenario, comparing it to several baseline approaches.

To effectively reduce JPEG compression artifacts, we propose a deep convolutional sparse coding (DCSC) network architecture. We design our DCSC in the framework of classic learned iterative shrinkage-threshold algorithm. To focus on recognizing and separating artifacts only, we sparsely code the feature maps instead of the raw image. The final de-blocked image is directly reconstructed from the coded features. We use dilated convolution to extract multi-scale image features, which allows our single model to simultaneously handle multiple JPEG compression levels. Since our method integrates model-based convolutional sparse coding with a learning-based deep neural network, the entire network structure is compact and more explainable. The resulting lightweight model generates comparable or better de-blocking results when compared with state-of-the-art methods.

Visual grounding, a task to ground (i.e., localize) natural language in images, essentially requires composite visual reasoning. However, existing methods over-simplify the composite nature of language into a monolithic sentence embedding or a coarse composition of subject-predicate-object triplet. In this paper, we propose to ground natural language in an intuitive, explainable, and composite fashion as it should be. In particular, we develop a novel modular network called Neural Module Tree network (NMTree) that regularizes the visual grounding along the dependency parsing tree of the sentence, where each node is a neural module that calculates visual attention according to its linguistic feature, and the grounding score is accumulated in a bottom-up direction where as needed. NMTree disentangles the visual grounding from the composite reasoning, allowing the former to only focus on primitive and easy-to-generalize patterns. To reduce the impact of parsing errors, we train the modules and their assembly end-to-end by using the Gumbel-Softmax approximation and its straight-through gradient estimator, accounting for the discrete nature of module assembly. Overall, the proposed NMTree consistently outperforms the state-of-the-arts on several benchmarks. Qualitative results show explainable grounding score calculation in great detail.

We propose a convolutional neural network (CNN) based image compression scheme that is compatible with JPEG-2000. Specifically, our scheme reuses the existing JPEG-2000 encoders to achieve bitstream, and features two components in addition to JPEG-2000: bitstream re-compression and decoder-side post-processing. First, we propose an advanced arithmetic codec that adopts CNN-based probability estimation to exploit the correlation between wavelet coefficients within and across subbands. Second, we propose a CNN-based post-processing method to improve the quality of reconstructed images. Experimental results show that the proposed two CNN-based components both help improve the compression efficiency by a significant margin.

Synaptic cleft is an important area for neuroscientists to analyze the macromolecular complexes related to neurotransmitter transmission. However, the large amount of noise and low signal-to-noise ratio in raw electron micrographs make it challenging to extract this region automatically. In this paper, we propose a simple but effective framework to automatically extract accurate boundaries of synaptic cleft regions. Our approach concatenates a novel contour growing algorithm to a fully convolutional network (FCN), so that it takes both advantages of large receptive field of FCNs and fine-level localization of contour evolution. The contour growing algorithm is based on the flexible evolving tension and synchronous growing controlling to localize the opening contour of clef region. With consideration of both global localization and local segmentation, our approach is more robust to noisy electron micrographs and outperforms all existing single-model FCNs on accurate segmentation of synaptic clefts.

Most micro-video recommender systems use the click-through to measure user satisfaction. However, the amount of time that users spend on a video, the playtime, measures user engagement on video contents and should be used as a complement to the click based signals. In this paper, we propose a coarse-to-fine multi-task jointly optimizing model to predict click-through and playtime. Following the click-through prediction, the playtime is first discretized into several intervals and classified into a specific one with a proposed ordered-balanced cross entropy loss. Then, to further improve upon coarse estimates, we learn a subtle offset with a regressor and produce a fine-grained playtime estimation. To make mutual promotion between click-through and playtime predictors, we optimize them jointly in a multi-task manner. Experimental results show that we achieve state-of-the-art performance on recommendation task and demonstrate effectiveness on playtime prediction at the same time.

Structure information is important to unsupervised depth learning from monocular videos. However, most existing methods focus on depth smoothing on planar regions, while other structure information, such as object shape and surface curvature, is ignored. In this work, we propose SGGN, a novel Structure Generation and Guidance Network to refine depth estimation under the guidance of extracted image structure. We introduce second-order Domain Transform filtering, which explores spatial depth variation by gradient propagation, to capture long-range dependence in the extracted structure for depth refinement. Then, several structure-aware constraints, as well as an attention mechanism, are applied to guide the training of SGGN, which leads to better depth estimation with structural guidance. Notably, our structure-aware constraints are designed in terms of different characteristics. Experiments on three benchmarks demonstrate the effectiveness of our structure-guided model and its state-of-the-art performance for unsupervised depth estimation.

A key problem in image set compression is inter-image prediction. Different from the conventional 2D transformation based methods, in this paper we propose a novel 3D mesh based inter-image prediction method. We reconstruct a 3D mesh from the images in the set as a compact representation of the photographed scene. Regarding the images as different projections of the mesh, we build coordinates mappings between images by the multi-view geometry. Exploiting the continuity of the mesh surface, we naturally model the occlusions in the scene and perform inter-image prediction with higher accuracy. The experimental results demonstrate that the proposed method outperforms the state-of-the-arts significantly.

Due to the fast inference and excellent learning capability, deep learning has become an effective means for image denoising and attracted considerable attention recently. However, for the images with rich textures and structures, the performance of deep learning approaches is still unsatisfactory. To address this issue, we develop a new convolutional neural network (CNN) for subband image denoising and name it SDCNN. In the proposed approach, we first decompose images into transform domain and denoise the coefficients of various subbands. By incorporating frequency information with spatial context, SDCNN is more effective in recovering image details. In particular, the introduced procedure of subband transform also plays the role of downsampling and enlarges the receptive field without increasing depth or sacrificing efficiency of network. Experimental results show that the SDCNN achieves promising results in terms of both objective and subjective performance.

Image paragraph generation aims to describe an image with a paragraph in natural language. Compared to image captioning with a single sentence, paragraph generation provides more expressive and fine-grained description for storytelling. Existing approaches mainly optimize paragraph generator towards minimizing word-wise cross entropy loss, which neglects linguistic hierarchy of paragraph and results in ``sparse" supervision for generator learning. In this paper, we propose a novel Densely Supervised Hierarchical Policy-Value (DHPV) network for effective paragraph generation. We design new hierarchical supervisions consisting of hierarchical rewards and values at both sentence and word levels. The joint exploration of hierarchical rewards and values provides dense supervision cues for learning effective paragraph generator. We propose a new hierarchical policy-value architecture which exploits compositionality at token-to-token and sentence-to-sentence levels simultaneously and can preserve the semantic and syntactic constituent integrity. Extensive experiments on the Stanford image-paragraph benchmark have demonstrated the effectiveness of the proposed DHPV approach with performance improvements over multiple state-of-the-art methods.

Non-orthogonal multiple access (NOMA) has shown potential for scalable multicast of video data. However, one key drawback for NOMA-based video multicast is the limited number of layers allowed by the embedded successive interference cancellation algorithm, failing to meet satisfaction of heterogeneous receivers. We propose a novel receiver-driven superposed video multicast (Supcast) scheme by integrating Softcast, an analog-like transmission scheme, into the NOMA-based system to achieve high bandwidth efficiency as well as gradual decoding quality proportional to channel conditions at receivers. Although Softcast allows gradual performance by directly transmitting power-scaled transformation coefficients of frames, it suffers performance degradation due to discarding coefficients under insufficient bandwidth and its power allocation strategy cannot be directly applied in NOMA due to interference. In Supcast, coefficients are grouped into chunks, which are basic units for power allocation and superposition scheduling. By bisecting chunks into base-layer chunks and enhanced-layer chunks, the joint power allocation and chunk scheduling is formulated as a distortion minimization problem. A two-stage power allocation strategy and a near-optimal low-complexity algorithm for chunk scheduling based on the matching theory are proposed. Simulation results have shown the advantage of Supcast against Softcast as well as the reference scheme in NOMA under various practical scenarios.

Entropy coding is a fundamental technique in video coding to remove the statistical redundancy in syntax elements. Currently, context-adaptive binary arithmetic coding (CABAC) is used as the entropy coding tool in HEVC. Considering that the manually designed binarization and context models are not flexible to estimate the probability of the syntax elements, we use neural networks to estimate the probability of the syntax elements, then the estimated probabilities together with the values of the syntax elements are fed into an arithmetic coding engine to fulfill entropy coding. In this paper, we focus on the syntax elements of inter prediction information that consists of merge flag, merge index, reference index, motion vector difference and motion vector prediction index in HEVC under low-delay P (LDP) setting. Compared with the previous work on neural network-based arithmetic coding for intra prediction modes and intra DC coefficients, there are three new characteristics in this paper. First, surrounding syntax elements are directly fed into the neural network without converting to reconstructed pixels. Second, unified neural networks are designed for different prediction block sizes. Finally, dependency among the syntax elements in current prediction unit is omitted to improve parallelism. Experimental results show that compared with HEVC, our proposed method achieves up to 0.5% and on average 0.3% BD-rate reduction in LDP configuration.

The semi-global matching algorithm is usually used for generating high-quality and real-time disparity maps in stereo vision systems. To reduce the hardware-resource consumption, we present a multi-stage pipeline hardware architecture with timesharing reuse for semi-global stereo matching. Combined with image down-sampling, jumping disparity, and a post processing, the presented architecture is used in a practical advanced driver-assistance system (ADAS), which is implemented on a Zynq-7 FPGA chip. The whole stereo matching architecture consumes 19,603 LUTs and 61.5 BRAM (36 KB), and the throughput is 2857 Million Disparity Estimation per second (MDE/S), which corresponds to a throughput of 31 fps when processing images with 1280∗960 resolution and 75 disparity levels.

Contrast enhancement and noise removal are coupled problems for low-light image enhancement. The existing Retinex based methods do not take the coupling relation into consideration, resulting in under or over-smoothing of the enhanced images. To address this issue, this paper presents a novel progressive Retinex framework, in which illumination and noise of low-light image are perceived in a mutually reinforced manner, leading to noise reduction low-light enhancement results. Specifically, two fully pointwise convolutional neural networks are devised to model the statistical regularities of ambient light and image noise respectively, and to leverage them as constraints to facilitate the mutual learning process. The proposed method not only suppresses the interference caused by the ambiguity between tiny textures and image noises, but also greatly improves the computational efficiency. Moreover, to solve the problem of insufficient training data, we propose an image synthesis strategy based on camera imaging model, which generates color images corrupted by illumination-dependent noises. Experimental results on both synthetic and real low-light images demonstrate the superiority of our proposed approaches against the State-Of-The-Art (SOTA) low-light enhancement methods.

In this paper, a Hierarchical Temporal Model (HTM) is proposed for the video captioning task, based on exploring the global and local temporal structure to better recognize fine-grained objects and actions. In our HTM, the encoder and decoder are hierarchically aligned according to different levels of features. The encoder applies two LSTM layers to construct temporal structures at both frame-level and object-level where the attention mechanism is applied to locate objects of interest, and the decoder uses corresponding LSTM layers to extract pivotal features from global to local through multi-level attention mechanism. Moreover, the local temporal structure is constructed implicitly from candidate object-oriented features under the guidance of global temporal-spatial representation, that could generate more accurate descriptions in handling shot-switching problems. Experiments on the widely used Microsoft Video Description Corpus (MSVD) and Charades datasets demonstrate the effectiveness of our proposed approach when compared to the state-of-the-art methods.

In this paper, we explore the task of generating photo-realistic face images from lines. Previous methods based on conditional generative adversarial networks (cGANs) have shown their power to generate visually plausible images when a conditional image and an output image share well-aligned structures. However, these models fail to synthesize face images with a whole set of well-defined structures, e.g. eyes, noses, mouths, etc., especially when the conditional line map lacks one or several parts. To address this problem, we propose a conditional self-attention generative adversarial network (CSAGAN). We introduce a conditional self-attention mechanism to cGANs to capture long-range dependencies between different regions in faces. We also build a multi-scale discriminator. The large-scale discriminator enforces the completeness of global structures and the small-scale discriminator encourages fine details, thereby enhancing the realism of generated face images. We evaluate the proposed model on the CelebA-HD dataset by two perceptual user studies and three quantitative metrics. The experiment results demonstrate that our method generates high-quality facial images while preserving facial structures. Our results outperform state-of-the-art methods both quantitatively and qualitatively.

Affective video content analysis has attracted a lot of attention recently. However, it faces various challenges such as the gap between intrinsic visual-aural features and spontaneous human emotional response, as well as ubiquitously existed label noise in affective annotations. Therefore, it is difficult to get useful supervision signals to learn well-generalized patterns responsible for eliciting affective impact. Observing that label uncertainty severely obstacles the progress of affective video content analysis, a deep denoising framework is proposed to infer true latent labels and annotation qualities from heavy label noise, fully utilizing the multimodal information contained in videos. Specifically, a quality embedding network is adopted in a multimodal fashion, and corresponding stochastic gradient descent (SGD) optimization objective is derived with variational inference and conditional independence assumption. To better reflect the effectiveness of affective models, new test sets are established based on the widely used LIRIS-ACCEDE dataset where the training database is kept unchanged, and a ranking-based evaluation metric is introduced accordingly. Experiments conducted on both the original LIRIS-ACCEDE test dataset and the refined one demonstrate the effectiveness of the proposed method.

We propose a novel hybrid imaging system to acquire 4D high-speed hyperspectral (HSHS) videos with high spatial and spectral resolution. The proposed system consists of two branches: one branch performs Nyquist sampling in the temporal dimension while integrating the whole spectrum, resulting in a high-frame-rate panchromatic video; the other branch performs compressive sampling in the spectral dimension with longer exposures, resulting in a low-frame-rate hyperspectral video. Owing to the high light throughput and complementary sampling, these two branches jointly provide reliable measurements for recovering the underlying HSHS video. Moreover, the panchromatic video can be used to learn an over-complete 3D dictionary to represent each band-wise video sparsely, thanks to the inherent structural similarity in the spectral dimension. Based on the joint measurements and the self-adaptive dictionary, we further propose a simultaneous spectral sparse (3S) model to reinforce the structural similarity across different bands and develop an efficient computational reconstruction algorithm to recover the HSHS video. Both simulation and hardware experiments validate the effectiveness of the proposed approach. To the best of our knowledge, this is the first time that hyperspectral videos can be acquired at a frame rate up to 100fps with commodity optical elements and under ordinary indoor illumination.

Recently, convolutional neural network (CNN)-based methods have achieved remarkable progress in image and video super-resolution, which inspires research on down-/up-sampling-based image and video coding using CNN. Instead of hand-crafted filters for up-sampling, trained CNN models are believed to be more capable of improving image quality, thus leading to coding gain. However, previous studies either concentrated on intra-frame coding or performed down- and up-sampling of entire frame. In this paper, we introduce block-level down- and up-sampling into inter-frame coding with the help of CNN. Specifically, each block in the P or B frame can either be compressed at the original resolution or down-sampled and compressed at low resolution and then, up-sampled by the trained CNN models. Such block-level adaptivity is flexible to cope with the spatially variant texture and motion characteristics. We further investigate how to enhance the capability of CNN-based up-sampling by utilizing reference frames and study how to train the CNN models by using encoded video sequences. We implement the proposed scheme onto the high efficiency video coding (HEVC) reference software and perform a comprehensive set of experiments to evaluate our methods. The experimental results show that our scheme achieves superior performance to the HEVC anchor, especially at low bit rates, leading to an average 3.8%, 2.6%, and 3.5% BD-rate reduction on the HEVC common test sequences under random-access, low-delay B, and low-delay P configurations, respectively. When tested on high-definition and ultrahigh-definition sequences, the average BD-rate exceeds 5%.

Inter prediction is a fundamental technology in video coding to remove the temporal redundancy between video frames. Traditionally, the reconstructed frames are directly put into a reference frame buffer to serve as references for inter prediction. Using multiple reference frames increases the accuracy of inter prediction, but also incurs a waste of memory of the buffer since the content of reference frames is highly similar. To address this problem, we propose to organize the references at clip level in addition to frame level, i.e. the reference buffer stores not only reference frames, but also reference clips that are cropped regions selected from the reconstructed frames. Using clip-level references, we can manage the reference content more economically, since the content of multiple reference frames is divided into the singular content of each frame as well as the repetitive content that appears in multiple frames. For the repetitive content, only one copy is stored in reference clips so as to avoid duplicate. Moreover, using reference clips also facilitates the bit-rate allocation among reference content, i.e. the quality of each clip can be decided adaptively to achieve the rate-distortion optimization. In this paper, we propose a complete video coding framework using reference clips, and investigate the problems including how to generate reference clips as either singular content clips or repetitive content clips, how to manage the clips, how to utilize the clips for inter prediction, and how to allocate bit-rate among clips, in a systematic manner. The proposed video coding framework is implemented upon the state-of-the-art video coding scheme, High Efficiency Video Coding (HEVC). Experimental results show that our scheme achieves on average 5.1% and 5.0% BD-rate reduction than the HEVC anchor, in low-delay B and low-delay P settings, respectively. We believe that reference clip opens up a new dimension for optimizing inter prediction in video coding, and thus is worthy of further study.

Fractional-pixel motion compensation (MC) improves the efficiency of inter prediction and has been utilized extensively in video coding standards. The traditional methods of fractional-pixel MC usually follow the approach of interpolation, i.e., they adopt different kinds of filters, either fixed or adaptive, to interpolate fractional-pixel values from integer-pixel values in a reference picture. Different from the interpolation approach, in this paper, we formulate the fractional-pixel MC as an inter-picture regression problem, which is to predict the pixel values of the current to-be-coded picture from the integer-pixel values of a reference picture, given a fractional-pixel motion vector that relates the two pictures. We then propose to adopt convolutional neural network (CNN) models to approach the regression problem, inspired by the recent advances of CNN. Accordingly, we propose fractional-pixel reference generation CNN (FRCNN) for both uni-directional and bi-directional MC in video coding. We further investigate how to train FRCNN by using encoded video sequences, and empirically study the effect of different training data and different CNN structures. Moreover, we propose to integrate FRCNN into the high efficiency video coding (HEVC) scheme, and perform a comprehensive set of experiments to evaluate the effectiveness of FRCNN. The experimental results show that our proposed FRCNN achieves on average 3.9%, 2.7%, and 1.3% bits saving compared with HEVC, under low-delay P, low-delay B, and random-access configurations, respectively.

We study the dual problem of image super-resolution (SR), which we term image compact-resolution (CR). Opposite to image SR that hallucinates a visually plausible high-resolution image given a low-resolution input, image CR provides a low-resolution version of a high-resolution image, such that the low-resolution version is both visually pleasing and as informative as possible compared to the high-resolution image. We propose a convolutional neural network (CNN) for image CR, namely, CNN-CR, inspired by the great success of CNN for image SR. Specifically, we translate the requirements of image CR into operable optimization targets for training CNN-CR: the visual quality of the compact resolved image is ensured by constraining its difference from a naively downsampled version and the information loss of image CR is measured by upsampling/super-resolving the compact-resolved image and comparing that to the original image. Accordingly, CNN-CR can be trained either separately or jointly with a CNN for image SR. We explore different training strategies as well as different network structures for CNN-CR. Our experimental results show that the proposed CNN-CR clearly outperforms simple bicubic downsampling and achieves on average 2.25 dB improvement in terms of the reconstruction quality on a large collection of natural images. We further investigate two applications of image CR, i.e., low-bit-rate image compression and image retargeting. Experimental results show that the proposed CNN-CR helps achieve significant bits saving than High Efficiency Video Coding when applied to image compression and produce visually pleasing results when applied to image retargeting.

Motion compensation with fractional motion vector has been widely utilized in the video coding standards. The fractional samples are usually generated by fractional interpolation filters. Traditional interpolation filters are usually designed based on the signal processing theory with the assumption of band-limited signal, which cannot effectively capture the non-stationary property of video content and cannot adapt to the variety of video quality. In this paper, we reveal an intuitive property of the fractional interpolation problem, named invertibility. That is, the fractional interpolation filters should not only generate fractional samples from integer samples but also recover the integer samples from the fractional samples in an invertible manner. We prove in theory that the invertibility in the spatial domain is equivalent to the constant magnitude in the Fourier transform domain. Driven by the invertibility, we then develop a learning-based method to solve the fractional interpolation problem. Inspired by the advances of convolutional neural network (CNN), we propose to establish an end-to-end scheme using CNN to train invertibility-driven interpolation filter (InvIF). Different from the previous learning-based methods, the proposed training scheme does not need hand-crafted "ground truth" of fractional samples. The proposed InvIF is integrated into high efficiency video coding (HEVC), and extensive experiments are conducted to verify its effectiveness. The experimental results show that the proposed method can achieve on average 4.7% and 3.6% BD-rate reduction compared with the HEVC anchor, under low-delay-B and random-access configurations, respectively.

Capturing images at high ISO modes will introduce much realistic noise, which is difficult to be removed by traditional denoising methods. In this paper, we propose a novel denoising method for high ISO JPEG images via deep fusion of collaborative and convolutional filtering. Collaborative filtering explores the non-local similarity of natural images, while convolutional filtering takes advantage of the large capacity of convolutional neural networks (CNNs) to infer noise from noisy images. We observe that the noise variance map of a high ISO JPEG image is spatial-dependent and has a Bayer-like pattern. Therefore, we introduce the Bayer pattern prior in our noise estimation and collaborative filtering stages. Since collaborative filtering is good at recovering repeatable structures and convolutional filtering is good at recovering irregular patterns and removing noise in flat regions, we propose to fuse the strengths of the two methods via deep CNN. The experimental results demonstrate that our method outperforms the state-of-the-art realistic noise removal methods for a wide variety of testing images in both subjective and objective measurements. In addition, we construct a dataset with noisy and clean image pairs for high ISO JPEG images to facilitate research on this topic.

Information centric networking (ICN) has been regarded as an ideal architecture for the next-generation network to handle users' increasing demand for content delivery with in-network cache. While making better use of network resources and providing better service delivery, an effective access control mechanism is needed due to the widely disseminated contents. However, in the existing solutions, making cache-enabled routers or content providers authenticate users' requests causes high computation overhead and unnecessary delay. Also, the straight-forward utilization of advanced encryption algorithms makes the system vulnerable to DoS attacks. Besides, privacy protection and service accountability are rarely taken into account in this scenario. In this paper, we propose SEAF, a secure, efficient, and accountable edge-based access control framework for ICN, in which authentication is performed at the network edge to block unauthorized requests at the very beginning. We adopt group signature to achieve anonymous authentication and use hash chain technique to reduce greatly the overhead when users make continuous requests for the same file. At the same time, we provide an efficient revocation method to make our framework more robust. Furthermore, the content providers can affirm the service amount received from the network and extract feedback information from the signatures and hash chains. By formal security analysis and the comparison with related works, we show that SEAF achieves the expected security goals and possesses more useful features. The experimental results also demonstrate that our design is efficient for routers and content providers and bring in only slight delay for users' content retrieval.

We propose the first privacy-preserving approach to address the privacy issues that arise in multi-agent planning problems modeled as a Dec-POMDP. Our solution is a distributed message-passing algorithm based on trials, where the agents' policies are optimized using the cross-entropy method. In our algorithm, the agents' private information is protected using a public-key homomorphic cryptosystem. We prove the correctness of our algorithm and analyze its complexity in terms of message passing and encryption/decryption operations. Furthermore, we analyze several privacy aspects of our algorithm and show that it can preserve the agent privacy of non-neighbors, model privacy, and decision privacy. Our experimental results on several common Dec-POMDP benchmark problems confirm the effectiveness of our approach.

This paper proposes an image upsampling scheme by introducing a new gradient field sharpening transform that converts the blurry gradient field of upsampled low-resolution (LR) image to a much sharper gradient field of original high-resolution (HR) image. Different from the existing methods that need to figure out the whole gradient profile structure and locate the edge points, we derive a new approach that sharpens the gradient field adaptively only based on the pixels in a small neighborhood. To maintain image contrast, image gradient is adaptively scaled to keep the integral of gradient field stable. Finally the HR image is reconstructed by fusing the LR image with the sharpened HR gradient field. Experimental results demonstrate that the proposed algorithm can generate more accurate gradient field and produce super-resolved images with better objective and visual qualities. Another advantage is that the proposed gradient sharpening transform is very fast and suitable for low-complexity applications.

Inter-Cell Interference Coordination (ICIC) is a promising way to improve energy efficiency in wireless networks, especially where small base stations are densely deployed. However, traditional optimization based ICIC schemes suffer from severe performance degradation with complex interference pattern. To address this issue, we propose a Deep Reinforcement Learning with Deterministic Policy and Target (DRL-DPT) framework for ICIC in wireless networks. DRL- DPT overcomes the main obstacles in applying reinforcement learning and deep learning in wireless networks, i.e. continuous state space, continuous action space and convergence. Firstly, a Deep Neural Network (DNN) is involved as the actor to obtain deterministic power control actions in continuous space. Then, to guarantee the convergence, an online training process is presented, which makes use of a dedicated reward function as the target rule and a policy gradient descent algorithm to adjust DNN weights. Experimental results show that the proposed DRL-DPT framework consistently outperforms existing schemes in terms of energy efficiency and throughput under different wireless interference scenarios. More specifically, it improves up to 15% of energy efficiency with faster convergence rate.

In recent years, neuromorphic computing systems based on memristive crossbar have provided a promising solution to enable acceleration of neural networks. Meanwhile, most of the neural networks used in realistic applications are often sparse. If such sparse neural network is directly implemented on a single memristive crossbar, it would result in inefficient hardware realizations. In this work, we propose 3D-FNC, a 3D floorplanning framework for neuromorphic computing systems in consideration of both crossbar utilization and design cost. 3D-FNC groups neurons that connect more common neurons into one cluster, where the optimal number of clusters is determined by L-method. As a result, the connections of a neural network can be effectively mapped to memristive crossbars or discrete synapses. Finally, a 3D floorplanning for memristive crossbars and neurons is developed to reduce area and wirelength cost. Experimental results show that 3D-FNC can achieve highly hardware-efficient designs, compared to state-of-the-art.

Inspired by the recent advances in deep learning, we propose a novel iterative belief propagation-convolutional neural network (BP-CNN) architecture to exploit noise correlation for channel decoding under correlated noise. The standard BP decoder is used to estimate the coded bits, followed by a CNN to remove the estimation errors of the BP decoder and obtain a more accurate estimation of the channel noise. Iterating between BP and CNN will gradually improve the decoding SNR and hence result in better decoding performance. To train a well-behaved CNN model, we define a new loss function which involves not only the accuracy of the noise estimation but also the normality test for the estimation errors, i.e., to measure how likely the estimation errors follow a Gaussian distribution. The introduction of the normality test to the CNN training shapes the residual noise distribution and further reduces the BER of the iterative decoding, compared to using the standard quadratic loss function. We carry out extensive experiments to analyze and verify the proposed framework.

Neuron segmentation is an important task for automatic analyses of brain images that are of huge volume. Previous methods for neuron segmentation rely on handcrafted image features, and have difficulty in coping with high-resolution, low signal-to-noise-ratio brain images. Convolutional neural network (CNN) has achieved remarkable success in natural image segmentation, but CNN requires accurately labeled data for training that are difficult to achieve on brain images of huge volume. In this paper, we present a weakly supervised learning strategy to deal with the inaccurate training data problem, and thus adopt 3D CNN to perform automatic soma segmentation from brain images. We test our method on our own collected mouse brain images that are of single-neuron resolution, and results show that 3D CNN-based method outperforms the traditional methods by a significant margin.

Hyperspectral video acquisition usually requires high complexity hardware and reconstruction algorithms. In this paper, we propose a low complexity CNN-based method for hyperspectral reconstruction from ubiquitous RGB videos, which effectively exploits the temporal redundancies within RGB videos and generates high-quality hyperspectral output. Specifically, given an RGB video, we first design an efficient motion compensation network to align the RGB frames and reduce the large motion. Then, we design a temporal-adaptive fusion network to exploit the inter-frame correlation. The fusion network has the ability to determine the optimum temporal dependency within successive frames, which further promotes the hyperspectral reconstruction fidelity. Preliminary experimental results validate the superior performance of the proposed method over previous learning-based methods. To the best of our knowledge, this is the first time that RGB videos are utilized for hyperspectral reconstruction through deep learning.

In the state-of-the-art video coding standard-High Efficiency Video Coding (HEVC), context-adaptive binary arithmetic coding (CABAC) is adopted as the entropy coding tool. In CABAC, the binarization processes are manually designed, and the context models are empirically crafted, both of which incur that the probability distribution of the syntax elements may not be estimated accurately, and restrict the coding efficiency. In this paper, we adopt a convolutional neural network-based arithmetic coding (CNNAC) strategy, and conduct studies on the coding of the DC coefficients for HEVC intra coding. Instead of manually designing binarization process and context model, we propose to directly estimate the probability distribution of the value of the DC coefficient using densely connected convolutional networks. The estimated probability together with the real DC coefficient are then input into a multi-level arithmetic codec to fulfill entropy coding. Simulation results show that our proposed CNNAC leads to on average 22.47% bits saving compared with CABAC for the bits of DC coefficients, which corresponds to 1.6% BD-rate reduction than the HEVC anchor.

Fractional-pixel interpolation has been widely used in the modern video coding standards to improve the accuracy of motion compensated prediction. Traditional interpolation filters are designed based on the signal processing theory. However, video signal is non-stationary, making the traditional methods less effective. In this paper, we reveal that the interpolation filter can not only generate the fractional pixels from the integer pixels, but also reconstruct the integer pixels from the fractional ones. This property is called invertibility. Inspired by the invertibility of fractional-pixel interpolation, we propose an end-to-end scheme based on convolutional neural network (CNN) to derive the invertible interpolation filter, termed CNNInvIF. CNNlnvIF does not need the “ground-truth” of fractional pixels for training. Experimental results show that the proposed CNNInvIF can achieve up to 4.6% and on average 2.2% BD-rate reduction than HEVC under the low-delay P configuration.

Information Centric Networking (ICN) has been regarded as an ideal architecture for the next-generation network to handle users' increasing demand for content delivery with in-network cache. While making better use of network resources and providing better delivery service, an effective access control mechanism is needed due to wide dissemination of contents. However, in the existing solutions, making cache-enabled routers or content providers authenticate users' requests causes high computation overhead and unnecessary delay. Also, straightforward utilization of advanced encryption algorithms increases the opportunities for DoS attacks. Besides, privacy protection and service accountability are rarely taken into account in this scenario. In this paper, we propose a secure, efficient, and accountable access control framework, called SEAF, for ICN, in which authentication is performed at the network edge to block unauthorized requests at the very beginning. We adopt group signature to achieve anonymous authentication, and use hash chain technique to greatly reduce the overhead when users make continuous requests for the same file. Furthermore, the content providers can affirm the service amount received from the network and extract feedback information from the signatures and hash chains. By formal security analysis and the comparison with related works, we show that SEAF achieves the expected security goals and possesses more useful features. The experimental results also demonstrate that our design is efficient for routers and content providers, and introduces only slight delay for users' content retrieval.

Inspired by the great success of convolutional neural network (CNN) in computer vision, we propose a CNN-based method to refine the motion compensation in video coding. First, we study a simple CNN-based motion compensation refinement (CNNMCR) scheme, where we train a CNN to refine the motion compensated prediction directly. Second, we consider to exploit the contextual information for the refinement, and propose a more powerful CNNMCR scheme, where the CNN utilizes not only the motion compensated prediction, but also the neighboring reconstructed region to refine the prediction. We integrate the simple CNNMCR and the CNNMCR schemes into the High Efficiency Video Coding (HEVC) framework. Experimental results show that both schemes achieve better compression performance than the HEVC anchor, leading to on average 1.8% and 2.3% BD-rate reduction, respectively, under low-delay P configuration. Furthermore, the combination of our proposed CNNMCR and the overlapped block motion compensation (OBMC) technique provides as high as 5.2% BD-rate reduction.

Many vision-language tasks can be reduced to the problem of sequence prediction for natural language output. In particular, recent advances in image captioning use deep reinforcement learning (RL) to alleviate the "exposure bias'' during training: ground-truth subsequence is exposed in every step prediction, which introduces bias in test when only predicted subsequence is seen. However, existing RL-based image captioning methods only focus on the language policy while not the visual policy (eg, visual attention), and thus fail to capture the visual context that are crucial for compositional reasoning such as visual relationships (\eg, "man riding horse'') and comparisons (eg. "smaller cat"). To fill the gap, we propose a Context-Aware Visual Policy network (CAVP) for sequence-level image captioning. At every time step, CAVP explicitly accounts for the previous visual attentions as the context, and then decides whether the context is helpful for the current word generation given the current visual attention. Compared against traditional visual attention that only fixes a single image region at every step, CAVP can attend to complex visual compositions over time. The whole image captioning model --- CAVP and its subsequent language policy network --- can be efficiently optimized end-to-end by using an actor-critic policy gradient method with respect to any caption evaluation metric. We demonstrate the effectiveness of CAVP by state-of-the-art performances on MS-COCO offline split and online server, using various metrics and sensible visualizations of qualitative visual context. The code is available at \urlhttps://github.com/daqingliu/CAVP

Depth estimation from monocular images is an ill-posed and inherently ambiguous problem. Recently, deep learning technique has been applied for monocular depth estimation seeking data-driven solutions. However, most existing methods focus on pursuing the minimization of average depth regression error at pixel level and neglect to encode the global layout of scene, resulting in layout-inconsistent depth map. This paper proposes a novel Layout-Aware Convolutional Neural Network (LA-Net) for accurate monocular depth estimation by simultaneously perceiving scene layout and local depth details. Specifically, a Spatial Layout Network (SL-Net) is proposed to learn a layout map representing the depth ordering between local patches. A Layout-Aware Depth Estimation Network (LDE-Net) is proposed to estimate pixel-level depth details using multi-scale layout maps as structural guidance, leading to layout-consistent depth map. A dense network module is used as the base network to learn effective visual details resorting to dense feed-forward connections. Moreover, we formulate an order-sensitive softmax loss to well constrain the ill-posed depth inferring problem. Extensive experiments on both indoor scene (NYUD-v2) and outdoor scene (Make3D) datasets have demonstrated that the proposed LA-Net outperforms the state-of-the-art methods and leads to faithful 3D projections.

Lossy compression of image and video yields visually annoying artifacts including blocking, blurring, ringing, etc., especially at low bit rates. In-loop filtering techniques can reduce these artifacts, improve quality, and achieve coding gain accordingly. In this paper, we present a convolutional neural network (CNN) based in-loop filter for High Efficiency Video Coding (HEVC). First, we design a new CNN structure that is composed of multiple Variable-filter-size Residue-learning blocks, namely VRCNN-ext, for artifact reduction. VRCNN-ext is trained by natural images as well as their compressed versions at different quality levels. Second, we investigate a new in-loop filter based on the trained VRCNN-ext models. Specifically, we observed that using VRCNN-ext directly on the inter pictures is not effective. To solve this problem, we further train a classifier to decide whether to use VRCNN-ext for each coding unit (CU). The classifier makes decision based on the compressed information, thus avoiding the overhead bits to control the on/off of the CNN-based filter at the CU level. Experimental results show that our scheme achieves significant bits saving than the HEVC anchor, leading to on average 9.2%, 9.6% and 7.4% BD-rate reduction on the HEVC test sequences, under all-intra, low-delay B and random-access configurations, respectively.

Motion estimation and motion compensation are fundamental in video coding to remove the temporal redundancy between video frames. The current video coding schemes usually adopt block-based motion estimation and compensation using simple translational or affine motion models, which cannot efficiently characterize complex motions in natural video signal. In this paper, we propose a frame extrapolation method for motion estimation and compensation. Specifically, based on the several previous frames, our method directly extrapolates the current frame using a trained deep network model. The deep network we adopted is a redesigned Video Coding oriented LAplacian Pyramid of Generative Adversarial Networks (VC-LAPGAN). The extrapolated frame is then used as an additional reference frame. Experimental results show that the VC-LAPGAN is capable in estimating and compensating for complex motions, and extrapolating frames with high visual quality. Using the VC-LAPGAN, our method achieves on average 2.0% BD-rate reduction than High Efficiency Video Coding (HEVC) under low-delay P configuration.

Inspired by the progress of image and video super-resolution (SR) achieved by convolutional neural network (CNN), we propose a CNN-based residue SR method for video coding. Different from the previous works that operate in the pixel domain, i.e. down- and up-sampling of image or video frame, we propose to perform down- and up-sampling in the residue domain. Specifically, for each block, we perform motion estimation and compensation to achieve residual signal at the original resolution, then we down-sample the residue and compress it at low resolution, and perform residue SR using a trained CNN model. We design a new CNN for residue SR with the help of the motion compensated prediction signal. We integrate the residue SR method into the High Efficiency Video Coding (HEVC) scheme, providing mode decision at the level of coding tree unit. Experimental results show that our method achieves on average 4.0% and 2.8% BD-rate reduction under low-delay P and low-delay B configurations, respectively.

Thermal cameras can capture images invariant to illumination conditions. However, thermal facial images are difficult to be recognized by human examiners. In this letter, an end-to-end framework, which consists of a generative network and a detector network, is proposed to translate thermal facial images into visible ones. The generative network aims at generating visible images given the thermal ones. The detector can locate important facial landmarks on visible faces and help the generative network to generate more realistic images that are easier to be recognized. As demonstrated in the experiments, the faces generated by our method have good visual quality and maintain identity preserving features.

We propose a hybrid digital-analog framework for wireless video transmission, which benefits from both the high distortion-power performance of digital systems and the graceful performance degradation of analog systems. The proposed framework models video frames as a parallel Gaussian source, which is separated into digital and analog parts through scalar quantization. It features entropy coding and channel coding in digital transmission and power scaling in analog transmission. The key challenge in this framework is how to allocate the constrained power and bandwidth resources between and among digital and analog components to achieve minimal distortion at the receiver. Given the worst-case channel signal-to-noise ratio, we are able to derive a closed-form expression of the overall distortion. However, minimizing it is a mixed-integer non-linear programming problem, which is generally non-deterministic polynomial-time hard. By making reasonable and justified simplifications, we approach the optimal solution through a practical scheme. Evaluations show that the proposed scheme outperforms the state-of-the-art analog scheme SoftCast by a large margin. The gain in received video peak signal-to-noise ratio is up to 5.0 dB for various types of videos.

Transform, as one of the most important modules of mainstream video coding systems, seems very stable over the past several decades. However, recent developments indicate that bringing more options for transform can lead to coding efficiency benefits. In this paper, we go further to investigate how the coding efficiency can be improved over the state-of-the-art method by adapting a transform for each block. We present a variable block-sized signal-dependent transforms (SDTs) design based on the High Efficiency Video Coding (HEVC) framework. For a coding block ranged from $4\times4$ to $32\times32$ , we collect a quantity of similar blocks from the reconstructed area and use them to derive the Karhunen-Loève transform. We avoid sending overhead bits to denote the transform by performing the same procedure at the decoder. In this way, the transform for every block is tailored according to its statistics, to be signal-dependent. To make the large block-sized SDTs feasible, we present a fast algorithm for transform derivation. Experimental results show the effectiveness of the SDTs for different block sizes, which leads to up to 23.3% bit-saving. On average, we achieve BD-rate saving of 2.2%, 2.4%, 3.3%, and 7.1% under AI-Main10, RA-Main10, RA-Main10, and LP-Main10 configurations, respectively, compared with the test model HM-12 of HEVC. The proposed scheme has also been adopted into the joint exploration test model for the exploration of potential future video coding standard.

The 3D extension of High Efficiency Video Coding significantly improves the coding efficiency of 3D video at the expense of computational complexity. This paper presents a novel fast mode decision algorithm for depth map coding based on the grayscale similarity and inter-view correlation. First, depth map grayscale similarity is adopted to judge whether the reference frame could assist the coding of the current frame. When the difference in the average grayscale between the co-located coding unit (CU) and the current CU is smaller than the similarity threshold, the depth level of the current CU will be restricted by that of the coded reference CU. Second, the grayscale similarity and inter-view correlation are jointly used for dependent views to achieve early decision on the best prediction unit (PU) mode. The mode decision procedure will be determined early when the co-located CU, which has a grayscale similarity with the current CU, selects Merge or Inter 2N ×2N as the best prediction mode. Moreover, when the corresponding CU in the independent view selects Merge or Inter 2N × 2N as the best prediction mode, the current CU will skip other PU modes checking based on the strong inter-view correlation. Finally, different strategies are proposed for the P-frames and B-frames of dependent views in view of the characteristics of different prediction structures. For B frames, the PU mode information of the coded independent view is utilized as reference to skip the unnecessary mode decision processes. For P frames, the spatial-temporal correlation is considered in the process of early mode decision to determine whether to choose the Merge mode or Inter 2N × 2N as the best mode. Experimental results show that our proposed scheme achieves considerable time saving with negligible degradation of coding performance.

In this paper, a novel rate-control algorithm based on the region adaptive R-λ model is proposed for depth maps coding. First, in order to obtain an accurate rate control for depth maps coding, a modified frame level bit allocation method based on coding bits statistical distribution of depth maps is proposed. Second, considering that different areas in a depth map have an imparity effect on virtual view rendering, the blocks of the depth map are divided into two types, namely, interested blocks for virtual view rending (IBV) and noninterested blocks for virtual view rending (NIBV). Then, two different R-λ models are derived for IBV and NIBV, respectively. The optimal bitrates for IBV and NIBV are determined by solving an optimization problem. After that, based on the regional R-λ models, the optimal Lagrange multipliers are calculated for both IBV and NIBV. Finally, the largest coding unit (LCU) level rate control is performed by adaptively adjusting the Lagrange multiplier to avoid blocking artifacts and smooth the quality of coding. Experimental results demonstrate that the proposed method can achieve considerable BD-PSNR gains compared with the unified rate-quantization model and conventional R-λ modelbased algorithms in terms of rendered virtual views quality.

Inspired by the recent advances of image super-resolution using convolutional neural network (CNN), we propose a CNN-based block up-sampling scheme for intra frame coding. A block can be down-sampled before being compressed by normal intra coding, and then up-sampled to its original resolution. Different from previous studies on down/up-sampling-based coding, the up-sampling methods in our scheme have been designed by training CNN instead of hand-crafted. We explore a new CNN structure for up-sampling, which features deconvolution of feature maps, multi-scale fusion, and residue learning, making the network both compact and efficient. We also design different networks for the up-sampling of luma and chroma components, respectively, where the chroma up-sampling CNN utilizes the luma information to boost its performance. In addition, we design a two-stage up-sampling process, the first stage being within the block-by-block coding loop, and the second stage being performed on the entire frame, so as to refine block boundaries. We also empirically study how to set the coding parameters of down-sampled blocks for pursuing the frame-level rate-distortion optimization. Our proposed scheme is implemented into the high-efficiency video coding (HEVC) reference software, and a comprehensive set of experiments have been performed to evaluate our methods. Experimental results show that our scheme achieves significant bits saving compared with the HEVC anchor, especially at low bit rates, leading to on average 5.5% BD-rate reduction on common test sequences and on average 9.0% BD-rate reduction on ultrahigh definition test sequences.

In this paper, we study a simplified affine motion model-based coding framework to overcome the limitation of a translational motion model and maintain low-computational complexity. The proposed framework mainly has three key contributions. First, we propose to reduce the number of affine motion parameters from 6 to 4. The proposed four-parameter affine motion model can not only handle most of the complex motions in natural videos, but also save the bits for two parameters. Second, to efficiently encode the affine motion parameters, we propose two motion prediction modes, i.e., an advanced affine motion vector prediction scheme combined with a gradient-based fast affine motion estimation algorithm and an affine model merge scheme, where the latter attempts to reuse the affine motion parameters (instead of the motion vectors) of neighboring blocks. Third, we propose two fast affine motion compensation algorithms. One is the one-step sub-pixel interpolation that reduces the computations of each interpolation. The other is the interpolation-precision-based adaptive block size motion compensation that performs motion compensation at the block level rather than the pixel level to reduce the number of interpolation. Our proposed techniques have been implemented based on the state-of-the-art high-efficiency video coding standard, and the experimental results show that the proposed techniques altogether achieve, on average, 11.1% and 19.3% bits saving for random access and low-delay configurations, respectively, on typical video sequences that have rich rotation or zooming motions. Meanwhile, the computational complexity increases of both the encoder and the decoder are within an acceptable range.

Analog-transmission-based soft video delivery suffers from the leveling-off effect when the allocated channel bandwidth is severely insufficient. Fortunately, with superimposed modulation, it is possible for analog traffic to share bandwidth with digital traffic. In this paper, we design and analyze such a hybrid digital-analog superimposed modulation (HDA-SIM) scheme for soft video delivery. Unlike previous work, we treat the bandwidth of competing digital traffic as hidden resources for the video delivery system. The key problem in this scheme is how to allocate the bandwidth and power resources among various modulation symbols so that we can improve the performance of video delivery without sacrificing the throughput of existing digital traffic. The resource allocation problem is formulated and the optimal solution under any given channel signal-to-noise ratio is derived. Based on the results, the sufficient and necessary condition for the video delivery system to achieve performance gain is given. In addition, we implement the proposed scheme for two state-of-the-art soft video delivery systems known as SoftCast and SharpCast. Both simulations and testbed evaluations show that the HDA-SIM version can achieve significant gains in the received video quality over their original designs.

In image restoration tasks, image priors generally utilize correlation within image contents to predict the latent image signal. In this paper, we propose to jointly exploit both intra- and inter-patch correlation of the input image, so as to further reduce the uncertainty of the unknown signal, and thus improve the prediction of the latent image. The proposed scheme evolves from the low-rank regularization for non-local highly-correlated image contents. Since the underlying cost function to pursue minimal rank is hard to solve, we use non-convex smooth surrogates for the rank penalty. Two such surrogates are utilized in order to incorporate both intra- and inter-patch correlation. To tackle the optimization problem, we use iterative alternating direction technique to divide the problem into two subproblems, each of which is solved via an empirical Bayesian procedure built upon variational approximation. Experimental results on image denoising show that the proposed approach outperforms several state-of-the-art methods in terms of peak signal-to-noise ratio, structural similarity, and perceptual quality.

This letter presents a novel approach for simultaneous depth and spectral imaging with a cross-modal stereo system. Two images of the target scene are captured at the same time: one compressively sampled hyperspectral measurement and one panchromatic measurement. The underlying hyperspectral cube is first reconstructed by leveraging the compressive sensing theory, during which a self-adaptive dictionary is learned from the panchromatic measurement to facilitate the reconstruction. The depth information of the scene is then recovered by estimating a disparity map between the hyperspectral cube and the panchromatic measurement through stereo matching. This disparity map, once obtained, is used to align the hyperspectral and panchromatic measurements to boost the hyperspectral reconstruction in an iterative manner. Through hardware experiments, for the first time to our knowledge, we demonstrate a snapshot system that allows for simultaneous depth and spectral imaging. The proposed system is capable of recording depth and spectral videos of dynamic scenes.

Unlike camera-captured video, screen content (SC) often contains a lot of repeating patterns, which makes some blocks used as references much more important than others. However, conventional rate-distortion optimization (RDO) schemes in video coding do not consider the dependence among image blocks, which often leads to a locally optimal parameter selection, especially for SC. In this paper, we present a weighted RDO scheme for SC coding (SCC), in which the repeating characteristics are taken into account when deciding RD tradeoff for each block. For one block, the number being referenced by the current picture and following pictures is estimated and based on the number, we set a proper weight in the RDO process to reflect its importance from a global point of view. To estimate the number being referenced, we propose a hash-based method to approximate the results to avoid the complexity of direct search. Experimental results show that compared with the High Efficiency Video Coding SCC reference software, 10.1%, 14.5%, and 2.2% on average and up to 25.7%, 39.8%, and 4.6% bit saving can be achieved by considering weights provided by our scheme for hierarchical-B, IBBB, and all intra coding structures, respectively. Thanks to our hash-based design, the complexity increase brought by the proposed scheme is marginal.

In the latest High Efficiency Video Coding (HEVC) development, i.e., HEVC screen content coding extensions (HEVC-SCC), a hash-based inter-motion search/block matching scheme is adopted in the reference test model, which brings significant coding gains to code screen content. However, the hash table generation itself may take up to half the encoding time and is thus too complex for practical usage. In this paper, we propose a hierarchical hash design and the corresponding block matching scheme to significantly reduce the complexity of hash-based block matching. The hierarchical structure in the proposed scheme allows large block calculation to use the results of small blocks. Thus, we avoid redundant computation among blocks with different sizes, which greatly reduces complexity without compromising coding efficiency. The experimental results show that compared with the hash-based block matching scheme in the HEVC-SCC test model (SCM)-6.0, the proposed scheme reduces about 77% of hash processing time, which leads to 12% and 16% encoding time savings in random access (RA) and low-delay B coding structures. The proposed scheme has been adopted into the latest SCM. A parallel implementation of the proposed hash table generation on graphics processing unit (GPU) is also presented to show the high parallelism of the proposed scheme, which achieves more than 30 frames/s for 1080p sequences and 60 frames/s for 720p sequences. With the fast hash-based block matching integrated into x265 and the hash table generated on GPU, the encoder can achieve 11.8% and 14.0% coding gains on average for RA and low-delay P coding structures, respectively, for real-time encoding.

This paper proposes a single-image super-resolution scheme by introducing a gradient field sharpening transform that converts the blurry gradient field of upsampled low-resolution (LR) image to a much sharper gradient field of original high-resolution (HR) image. Different from the existing methods that need to figure out the whole gradient profile structure and locate the edge points, we derive a new approach that sharpens the gradient field adaptively only based on the pixels in a small neighborhood. To maintain image contrast, image gradient is adaptively scaled to keep the integral of gradient field stable. Finally, the HR image is reconstructed by fusing the LR image with the sharpened HR gradient field. Experimental results demonstrate that the proposed algorithm can generate more accurate gradient field and produce super-resolved images with better objective and visual qualities. Another advantage is that the proposed gradient sharpening transform is very fast and suitable for low-complexity applications.

Hybrid digital-analog (HDA) transmission is becoming an attractive solution for mobile video delivery because it not only has graceful degradation with channel variations but also yields high power efficiency. However, the heavy bandwidth demand of analog transmission is still an unsolved problem, limiting the received video quality when the bandwidth is not sufficient. To address this problem, we propose adopting Shannon-Kotel'nikov (SK) mapping for HDA video transmission and design an HDA scheme called SK-Cast. SK-Cast consists of a digital and an analog branch. In the digital branch, SK-Cast compresses the video sequence using an high efficiency video coding digital encoder to produce a base layer. The base layer is transmitted through digital methods with strong protection. The residual signals are then decorrelated using three-dimensional discrete cosine transform transform. The SK mapping is exploited to transmit these coefficients, as they can achieve efficient bandwidth compression. We address the resource allocation problems in SK-Cast, including the allocation between digital and analog branches and the allocation among analog symbols. The simulation results show that the SK-Cast outperforms the state-of-the-art HDA systems, including WSVC and SharpCast, and a digital scalable video coding system.

The rate-distortion in conventional digital systems is replaced by cost distortion in pseudo-analog systems where the cost consists of power and bandwidth. In this paper, we formulate the cost-distortion optimization problem in terms of a power-bandwidth pair versus distortion to bring an insight to pseudo-analog transmission. Using a divide-and-conquer strategy, the 3-D optimization problem of a power-bandwidth pair versus distortion is decomposed into two subproblems: power distortion and bandwidth distortion optimization. To solve the integer nonlinear optimization problem, we propose two prediction models that transform the partial summation of variances and the square roots of variances into continuous functions. The proposed models are used to derive the closed-form solutions for both optimization subproblems, and a tradeoff between power and bandwidth is discussed. Accordingly, the resource control algorithm is designed to allocate the fewest resources required to obtain specific video quality; power can be traded for bandwidth and vice versa. Our experimental results show that the proposed optimization models achieve stable perceptual video quality comparable to that of Groups of Pictures and use resources more efficiently than do SoftCast.

Redundancy is necessary for a storage system to achieve reliability. Frequent errors in large-scale storage systems, for example, cloud, make it desirable to reduce the cost of recovery. Among all types of data in cloud storage, videos generally occupy significant amounts of space due to high volumes and the rapid development of video sharing and video-on-demand services. Unlike general data, videos can tolerate a certain level of quality degradation. This paper investigates multilayer video representations, such as scalable videos and simulcast streaming, and proposes an unequal error protection scheme based on local reconstruction codes (LRC) for video storage. By providing less protection for less important layers or video copies, a better tradeoff between storage and repair cost is achieved. Both theoretical and simulation results show that such a tradeoff can be achieved over the LRC with equal error protection, though the recovered video quality might be slightly lower in rare cases.

This paper presents the first snapshot hyperspectral light field imager in practice. Specifically, we design a novel hybrid camera system to obtain two complementary measurements that sample the angular and spectral dimensions respectively. To recover the full 5D hyperspectral light field from the severely undersampled measurements, we then propose an efficient computational reconstruction algorithm by exploiting the large correlations across the angular and spectral dimensions through self-learned dictionaries. Simulation on an elaborate hyperspectral light field dataset validates the effectiveness of the proposed approach. Hardware experimental results demonstrate that, for the first time to our knowledge, a 5D hyperspectral light field containing 9x9 angular views and 27 spectral bands can be acquired in a single shot.

Effective capacity analysis is a powerful tool to investigate the impact of physical layer designs on the link layer delay-sensitive QoS performance, which is important for real-time multimedia applications. In this paper, we rigorously analyze the effective capacities of downlink large-scale antenna systems. The main focus is to establish the fundamental effective capacity in a very-large MIMO system, and to characterize the performance difference between co-located and distributed antenna layouts. To that end, we first analytically derive the closed-form effective capacities for two widely used linear precoding schemes, conjugate beamforming and zero-forcing beamforming. We then analyze the asymptotic average effective capacities when the number of BS antennas and the number of users grow unboundedly with a fixed ratio. The effective capacity gain of the distributed antenna layout over the co-located layout is established via theoretical analysis.

In this paper, we investigate a convolutional neural network (CNN) approach for light field (LF) super-resolution (SR). We are motivated by the assumption that image priors can be embedded into CNN, and both external and internal correlations are important in LFSR. The LF images are indeed natural images except for its angular resolution, so the external correlations help to super-resolve a single image from a collection of general images, whilst the internal correlations are essential to enhance a single view in LF with the details in the other views. Accordingly, we propose a two-stage CNN, where the two stages exploit the external and internal correlations, respectively. Moreover, to improve the generalization ability of the second-stage CNN for inter-view SR, we propose to align different views at patch level to compensate for the disparity that is essential to LFSR, thus the second stage is termed multi-patch fusion CNN. Experimental results demonstrate the superior performance of our two-stage CNN compared with the state-of-the-art CNN-based SR methods.

Inter prediction in video coding is very efficient to remove temporal redundancy. However, due to the limitation of short-term references, inter prediction can work only within a very short time interval. In surveillance videos, we observe that there are always similar vehicles passing through one static camera, but the time intervals of similar vehicles are usually several seconds to minutes, exceeding the time interval that short-term references can handle. To solve this problem, we propose to build a vehicle library, and to put high-quality copies of the similar vehicles into the vehicle library. During encoding, vehicles are detected from the current frame, and for each vehicle we can retrieve similar vehicles from the vehicle library, and take the retrieved vehicle picture as additional references for inter prediction. Preliminary experimental results show that the proposed vehicle library based method achieves as high as 10.1% bit-rate saving for surveillance video coding, compared to HEVC anchor.

Virtual reality (VR) has been a hot topic in both research and industry, calling for more efficient compression of panoramic videos. Currently, panoramic video is played as if it is spherical, but such video is actually mapped to planar video, e.g. using equirectangular projection, before compression and transmission. The projection causes deformation and thus makes the traditional translational motion model not efficient. In this paper, we propose a new motion model based on spherical coordinates transform to compensate for the deformation in panoramic videos. Our model requires no additional motion vector but rather derives pixel-wise 2D motion vectors from a block-level 3D motion vector. Our experimental results show the significant bits saving achieved by the new model, which leads to as high as 8.0% BD-rate on the test sequences.

This paper presents a novel super-resolution method for light field images by jointly exploiting internal and external similarities. The internal similarity refers to the correlations that exist across the angular dimensions of the 4D light field itself, while the external similarity refers to the correlations learned from a conventional 2D image dataset. Our key observation is that the internal and external similarities are complementary to each other, and we propose a depth-adaptive fusion scheme to take advantage of both their merits. Moreover, we improve the traditional projection-based method that exploits the internal similarity, by introducing a back-projection refinement and getting rid of the dependency on camera parameters. Experimental results on a variety of light field images validate the superior performance of the proposed method.

As the latest 3D video coding standard, 3D High Efficiency Video Coding (3D-HEVC) achieves efficient coding. Depth coding plays an important role in 3D-HEVC. For better prediction of edges in depth maps, new depth intra modes were proposed, such as Depth Modeling Modes (DMMs). DMM Mode 1, namely explicit wedgelet signalization mode, can improve the performance of synthesized views, but leads to unaffordable encoding computation complexity. In this paper, a simplified search algorithm for explicit wedgelet signalization mode is proposed to predigest the complex search process. First, the proposed algorithm only searches the partition patterns in a limited set based on edge detection in the coarse search stage, since the separation line in the best matching pattern should be similar to the edge in an actual depth prediction unit (PU). Then, a wide range of refinement is implemented to guarantee the coding performance. In fact, up to 24 refinements are tested to provide accurate prediction in the refinement process. Experimental results show that considerable encoding time saving is achieved with negligible performance loss. For all intra test cases, the proposed algorithm achieves an average encoding time saving of 75% for DMM Mode 1 with negligible bitrate increase on synthesized views, compared with the default coarse-refinement algorithm in HTM.

This paper addresses the challenge of 3-D skeleton recovery by exploiting the spatio-temporal correlations of corrupted 3D skeleton sequences. A skeleton sequence is represented as a matrix. We propose a novel low-rank solution that effectively integrates both a low-rank model for robust skeleton recovery based on temporal coherence, and an articulation-graph-based isometric constraint for spatial coherence, namely consistency of bone lengths. The proposed model is formulated as a constrained optimization problem, which is efficiently solved by the Augmented Lagrangian Method with a Gauss-Newton solver for the subproblem of isometric optimization. Experimental results on the CMU motion capture dataset and a Kinect dataset show that the proposed approach achieves better recovery accuracy over a state-of-the-art method. The proposed method has wide applicability for skeleton tracking devices, such as the Kinect, because these devices cannot provide accurate reconstructions of complex motions, especially in the presence of occlusion.

In this paper, we propose a novel integrated task planning system for service robot in domestic domains. Given open-ended high-level user instructions in natural language, robots need to generate a plan, i.e., a sequence of low-level executable actions, to complete the required tasks. To address this, we exploit the knowledge on semantic roles of common verbs defined in semantic dictionaries such as FrameNet and integrate it with Answer Set Programming --- a task planning framework with both representation language and solvers. In the experiments, we evaluated our approach using common benchmarks on service tasks and showed that it can successfully handle much more tasks than the state-of-the-art solution. Notably, we deployed the proposed planning system on our service robot for the annual RoboCup@Home competitions and achieved very encouraging results.

We propose a novel baseline regret minimization algorithm for multi-agent planning problems modeled as finite-horizon decentralized POMDPs. It guarantees to produce a policy that is provably better than or at least equivalent to the baseline policy. We also propose an iterative belief generation algorithm to effectively and efficiently minimize the baseline regret, which only requires necessary iterations to converge to the policy with minimum baseline regret. Experimental results on common benchmark problems confirm its advantage comparing to the state-of-the-art approaches.

Motion compensation is a fundamental technology in video coding to remove the temporal redundancy between video frames. To further improve the coding efficiency, sub-pel motion compensation has been utilized, which requires interpolation of fractional samples. The video coding standards usually adopt fixed interpolation filters that are derived from the signal processing theory. However, as video signal is not stationary, the fixed interpolation filters may turn out less efficient. Inspired by the great success of convolutional neural network (CNN) in computer vision, we propose to design a CNN-based interpolation filter (CNNIF) for video coding. Different from previous studies, one difficulty for training CNNIF is the lack of ground-truth since the fractional samples are actually not available. Our solution for this problem is to derive the “ground-truth” of fractional samples by smoothing high-resolution images, which is verified to be effective by the conducted experiments. Compared to the fixed half-pel interpolation filter for luma in High Efficiency Video Coding (HEVC), our proposed CNNIF achieves up to 3.2% and on average 0.9% BD-rate reduction under low-delay P configuration.

An increasing number of deployed surveillance cameras raise a huge demand for higher efficiency video coding scheme, and the emerging background reference based video coding methods with High Efficiency Video Coding (HEVC) have achieved a large increase of coding efficiency on surveillance videos. However, the high encoding complexity of HEVC causes troubles for these methods to be adopted in practice, especially in real-time coding scenarios. Among all the factors resulting in the increase of encoding complexity, mode decision and motion estimation are both critical reasons. Therefore, a fast algorithm based on a block-level background generation method is proposed to address this problem for surveillance video. With the help of fast detection of background coding units, the Merge mode is early decided, some prediction modes at specific depths that have the least probabilities are skipped using early termination during rate-distortion optimization, and unnecessary motion estimation is also avoided. Thanks to the full utilization of the static camera premise, the proposed fast algorithm achieves 77.0% reduction of encoding time, while only incurs 1.0% performance loss that is negligible. More experimental results also verify that the proposed algorithm outperforms the state-of-the-art schemes.

Based on the observation that a matrix X consisted of non-local highly-correlated patches is of low rank, many image restoration methods use low-rank regularization to exploit correlation between image contents, so that the uncertainty of the unknown image signal can be reduced. To tackle the problem that the underlying cost function to pursue minimal rank is hard to solve, an effective way is to employ smooth non-convex surrogate log |XXT| for the rank penalty. Essentially, such technique only considers to utilize correlation within image patches. In this paper, we propose to jointly exploit both intra- and inter-patch correlation of the input image, so as to further reduce the uncertainty of the signal, and thus improve the prediction of the latent image. The corresponding two surrogates are integrated to incorporate both intra- and inter-patch correlation. To solve the optimization problem, we use iterative alternating direction technique to divide the problem into two subproblems, each of which is solved via an empirical Bayesian procedure built upon variational approximation. Experimental results show that the proposed approach outperforms several state-of-the-art methods in terms of PSNR and perceptual quality.

This paper combines an adaptive reconstruction based approach and a learning based technique into an effective scheme for single image super-resolution. Unlike conventional schemes that adopt pre-trained dictionaries to tell the relationship between high-resolution (HR) image and the low-resolution (LR) observation, the proposed method attempts to learn the joint distribution of highly-correlated patch couples from the input image itself instead of an external dataset, so that the learnt models are specially tailored for the current patches and thus can better fit the image data of interest. To be specific, we first apply spatially adaptive gradient sparsity regularization in the reconstruction of the HR image using the contour information, and then utilize the generated HR output to guide the joint distribution learning to infer the relationship between the highly-correlated HR and LR patches. In this way, we simultaneously exploit the inter-scale correlation as well as the local and non-local correlation of the image contents. Empirical results show that the performance of the proposed method is highly competitive with state-of-the-art schemes in terms of peak signal-to-noise ratio (PSNR) and perceptual quality.

Light field (LF) cameras offer the capability of depth estimation in a single shot, which facilitates real-time 3D reconstruction of dynamic scenes. However, the accuracy of depth estimated from LF is still limited. Different from previous methods that generally focus on improving the fidelity of the central view depth, we argue that depth maps obtained at different views contain complementary information. Inspired by the principle of Kinect-fusion, we then propose a novel method for dense and accurate 3D reconstruction from LF images, namely, LF-fusion. Specifically, we use the iterative closest point (ICP) algorithm to register the point clouds generated from different views, and then employ a volumetric integration algorithm based on the truncated signed distance function (TSDF) to reconstruct the final 3D surface. Experiments demonstrate that the proposed method produces superior 3D reconstruction results on two representative LF datasets.

In both H.264 and HEVC, context-adaptive binary arithmetic coding (CABAC) is adopted as the entropy coding method. CABAC relies on manually designed binarization processes as well as handcrafted context models, which may restrict the compression efficiency. In this paper, we propose an arithmetic coding strategy by training neural networks, and make preliminary studies on coding of the intra prediction modes in HEVC. Instead of binarization, we propose to directly estimate the probability distribution of the 35 intra prediction modes with the adoption of a multi-level arithmetic codec. Instead of handcrafted context models, we utilize convolutional neural network (CNN) to perform the probability estimation. Simulation results show that our proposed arithmetic coding leads to as high as 9.9% bits saving compared with CABAC.

Recently, a scheme called SoftCast has shown great potential for wireless image/video communication in the scenarios where the channel quality may fluctuate drastically and unpredictably. The transmission is lossy in nature, with its transmission power allocated among coefficients unequally to minimize the distortion. One problem is that its performance is optimized using mean square errors (MSE) as the quality metric, which is known for not matching the perception of human eyes. Inspired by the researches in image quality assessment, this paper proposes a power allocation and optimization scheme that minimizes the perceptual distortion of reconstruction image. In particular, we establish a perception model to evaluate the perceptual importance of different transform coefficients, based on the structure similarity (SSIM) image quality metric. Experimental results show that the proposed scheme can improve the perceptual performance of the original SoftCast scheme.

Leveraging the compressive sensing (CS) theory, coded aperture snapshot spectral imaging (CASSI) provides an efficient solution to recover 3D hyperspectral data from a 2D measurement. The dual-camera design of CASSI, by adding an uncoded panchromatic measurement, enhances the reconstruction fidelity while maintaining the snapshot advantage. In this paper, we propose an adaptive nonlocal sparse representation (ANSR) model to boost the performance of dual-camera compressive hyperspectral imaging (DCCHI). Specifically, the CS reconstruction problem is formulated as a 3D cube based sparse representation to make full use of the nonlocal similarity in both the spatial and spectral domains. Our key observation is that, the panchromatic image, besides playing the role of direct measurement, can be further exploited to help the nonlocal similarity estimation. Therefore, we design a joint similarity metric by adaptively combining the internal similarity within the reconstructed hyperspectral image and the external similarity within the panchromatic image. In this way, the fidelity of CS reconstruction is greatly enhanced. Both simulation and hardware experimental results show significant improvement of the proposed method over the state-of-the-art.

A block-composed background reference method is proposed in this paper for High Efficiency Video Coding (HEVC). For a group of picture (GoP), the first reconstructed picture is served as an initial background reference, which probably includes foreground content. In the subsequent coding, some background coding tree units (CTUs) in every picture are selected to be compressed with high quality. These reconstructed CTUs are used to update the background reference as well as replace the foreground content. Finally, a high-quality background reference is generated to better exploit the long-term temporal correlation in the video. There are three key technical contributions in the proposed coding scheme. First, the background reference is generated gradually by block updating instead of picture updating, which makes the scheme free of bit-rate burst and more suitable for real-time applications and can generate high-quality background reference even with complicated foreground. Second, we propose an approach to select background CTUs by taking both temporal and spatial smoothness into account. Third, we propose a model to decide the coding parameters of the selected background CTUs based on the overall picture activity, which essentially pursues the GoP-level optimal performance when making CTU-level decision. The proposed background reference is implemented into HEVC, and the experimental results demonstrate a significant improvement in coding efficiency. Compared with HEVC, our method can averagely save 14% bits in surveillance and conferencing sequences with negligible increase of encoding and decoding complexity. In particular, it can still averagely save 7.3% bits in HEVC general test sequences. Obviously, the proposed scheme can be applied to more general video contents.

Accurate and high-quality depth maps are required in lots of 3D applications, such as multi-view rendering, 3D reconstruction and 3DTV. However, the resolution of captured depth image is much lower than that of its corresponding color image, which affects its application performance. In this paper, we propose a novel depth map super-resolution (SR) method by taking view synthesis quality into account. The proposed approach mainly includes two technical contributions. First, since the captured low-resolution (LR) depth map may be corrupted by noise and occlusion, we propose a credibility based multi-view depth maps fusion strategy, which considers the view synthesis quality and interview correlation, to refine the LR depth map. Second, we propose a view synthesis quality based trilateral depth-map up-sampling method, which considers depth smoothness, texture similarity and view synthesis quality in the up-sampling filter. Experimental results demonstrate that the proposed method outperforms state-of-the-art depth SR methods for both super-resolved depth maps and synthesized views. Furthermore, the proposed method is robust to noise and achieves promising results under noise-corruption conditions.

Recently, there is a resurgence of interest in uncoded transmission for wireless visual communication. While conventional coded systems suffer from cliff effect as the channel condition varies dynamically, uncoded linear-transformed transmission (ULT) provides elegant quality degradation for wide channel SNR range. ULT skips non-linear operations, such as quantization and entropy coding. Instead, it utilizes linear decorrelation transform and linear scaling power allocation to achieve optimized transmission. This paper presents a theoretical analysis for power-distortion optimization of ULT. In addition to the observation in our previous work that a decorrelation transform can bring significant performance gain, this paper reveals that exploiting the energy diversity in transformed signal is the key to achieve the full potential of decorrelation transform. In particular, we investigated the efficiency of ULT with exact or inexact signal statistics, highlighting the impact of signal energy modeling accuracy. Based on that, we further proposed two practical energy modeling schemes for ULT of visual signals. Experimental results show that the proposed schemes improve the quality of reconstructed images by 3~5 dB, while reducing the signal modeling overhead from hundreds or thousands of meta data to only a few meta data. The perceptual quality of reconstruction is significantly improved.

Moiré artifacts are generally caused by the interference between the overlap of the sensor's sampling grid and high-frequency (nearly) periodic textures, and heavily affect the image quality. However, it is difficult to effectively remove moiré artifacts from textured images as the structure of moiré patterns is similar to that of textures in some sense. In this paper, we propose a novel textured image demoiréing method by signal decomposition and guided filtering. Given a textured image with moiré artifacts, we first remove moiré artifacts in the green (G) channel using the proposed low-rank and sparse matrix decomposition model. This model regularizes the texture layer by the low-rank prior in spatial domain and the moiré layer by sparse representation in frequency domain. An alternating direction method under the augmented Lagrangian multiplier framework is used to solve the matrix decomposition model. Then, since the red (R) and blue (B) channels are more heavily polluted by moiré artifacts than the G channel, we propose to remove moiré artifacts in its R and B channels via guided filtering by the obtained texture layer of the G channel. Experimental results demonstrate that our method outperforms the state-of-the-art methods for both synthetic and real images.

In this paper, we propose to introduce intrinsic image decomposition priors into decomposition models for contrast enhancement. Since image decomposition is a highly illposed problem, we introduce constraints on both reflectance and illumination layers to yield a highly reliable solution. We regularize the reflectance layer to be piecewise constant by introducing a weighted ℓ1norm constraint on neighboring pixels according to the color similarity, so that the decomposed reflectance would not be affected much by the illumination information. The illumination layer is regularized by a piecewise smoothness constraint. The proposed model is effectively solved by the Split Bregman algorithm. Then, by adjusting the illumination layer, we obtain the enhancement result. To avoid potential color artifacts introduced by illumination adjusting and reduce computing complexity, the proposed decomposition model is performed on the value channel in HSV space. Experiment results demonstrate that the proposed method performs well for a wide variety of images, and achieves better or comparable subjective and objective quality compared with the state-of-the-art methods.

We propose a progressive pseudo-analog video transmission scheme that simultaneously handles SNR and bandwidth variations with graceful quality degradation for mobile video streaming. With the inherited SNR-adaptability from pseudo-analog transmission, the proposed progressive solution acquires bandwidth adaptability through an innovative scheduling algorithm with optimal power allocation. The basic idea is to aggressively transmit or retransmit important coefficients so that distortion is minimized at the receiver after each received packet. We derive the closed-form expression of reduced distortion for each packet under given transmission power and known channel conditions, and show that the optimal solution can be obtained with a water-filling algorithm. We also illustrate through analyses and simulations that a near-optimal solution can be found through approximation when only statistical channel information is available. Simulations show that our solution approaches the performance upper bound of pseudo-analog transmission in an additive white Gaussian noise channel and significantly outperforms existing pseudo-analog solutions in a fast Rayleigh fading channel. Trace-driven emulations are also carried out to demonstrate the advantage of the proposed solution over the state-of-the-art digital and pseudo-analog solutions under a real dramatically varying wireless environment.

In this paper, we propose a pixel fusion-based stereo image retargeting method, which could adaptively retarget stereo images with flexible aspect ratios, simultaneously preserving the depth. Retargeting each image independently by the pixel fusion method ignores the disparity relationship between pixels in the image pair and hence will introduce the distortion of disparity. To address this issue, we advocate to extend the single pixel fusion-based way to be applicable for stereo image pair. First, seams are selected based on the energy function, which simultaneously considers the seam selecting and seam matching. Second, a seam-matching-based matching map is proposed to preserve the disparity relationship between image pair. Then, the scaling factors for the left image are assigned considering both the important object and depth preservation. Subsequently, the scaling factors for the right image are obtained according to the proposed matching map. Based on these scaling factors, the stereo image pair is retargeted with pixel fusion. In contrast to removing pixels to resize image, the way of pixel fusion can obtain more smooth results with less depth distortion. Experimental results demonstrate that our method achieves more preferable qualities in both depth and shape preservation for stereo image retargeting.

We consider efficient image transmission via time-varying channels. To improve the performance, we propose a new distributed compressive sensing (CS) scheme that can leverage similar images in the cloud. It is featured by channel SNR and bandwidth scalability, high efficiency, and low encoding complexity. For each image, a compressed thumbnail is first transmitted after forward error correction (FEC) and modulation to retrieve similar images and generate a side information (SI) in the cloud. The residual image after subtracting the decompressed thumbnail is then coded and transmitted by CS through a very dense constellation without FEC. The linearly and ratelessly generated CS measurements make it capable of achieving both graceful quality degradation (GD) with the channel SNR and bandwidth scalability in a universal scheme. A mode decision and transform-domain power allocation are introduced for better bandwidth usage and protection against channel errors. At the decoder, a two-step CS decoding is performed to recover the residual signal, where both the local and nonlocal correlations within the image and that with the SI are exploited. Simulations on landmark images and an AWGN channel show that the received image quality gracefully increases with the channel SNR and bandwidth. Furthermore, it outperforms existing schemes both subjectively and objectively by up to 11 dB gains compared with the state-of-the-art transmission scheme with GD, i.e. SoftCast.

The background information is a significant prior for salient object detection, especially when images contain cluttered background and diverse object parts. In this paper, we propose a background-driven salient object detection (BD-SOD) method to more comprehensively exploit the background prior, aiming at generating more accurate and robust salient maps. To be specific, we first exploit the background prior to conduct the saliency estimation, i.e., computing the regional saliency values. In this stage, the background prior is utilized in threefold: restricting the reference regions to only the background regions, weighting the contribution of reference regions, and leveraging the importance of different features. Benefiting from such an explicit utilization, the proposed model can greatly mitigate the negative interference of the cluttered background and diverse object parts. We then embed the background prior into the optimization graph for saliency refinement. Specifically, two virtual supernodes (representing the background and foreground, respectively) are introduced with extra connections, and the nonlocal feature connections between similar regions are also set up. These connections enhance the power of optimization graph to alleviate the perturbations from diverse parts, and thus help to achieve the uniformity of saliency values. Finally, we provide systematical studies to investigate the effectiveness of the proposed BD-SOD in exploiting the valuable background prior. Experimental results on multiple public benchmark datasets, including MSRA-1000, THUS-10000, PASCAL-S, and ECSSD, clearly show that BD-SOD consistently outperforms the well-established baselines and achieves state-of-the-art performance.

Different from conventional profiles in the state-of-the-art video coding standard HEVC and related optimization methods, we focus on building the optimally scalable algorithms under constrained and varying computational capacity to take full advantages of HEVC as far as possible in order to meet the growing demands of computational capacity adaptive applications such as real-time video communication and video coding on different mobile devices. We propose a video coding framework based on priority order for a special profile and give the general thoughts of designing algorithms by utilizing cost-performance as priority. For the framework, we invent a feasible solution by introducing a three-level coding structure and some novel features that express the relationship between video contents and their priorities. Experimental results partially prove our framework may nearly achieve the optimal coding efficiency under the continuously changeable computing limitations at every time with negligible extra time consuming.

The directional intra prediction (DIP) modes in HEVC are capable of predicting local continuous image features. Recently, intra block copy (IBC) is proposed for screen content coding, aiming at predicting non-local recurrent image features. For natural video, we observe that recurrent features are often irregular and not aligned with blocks. Thus, we propose a combination of DIP and IBC with block partition for better intra prediction, where one block can be divided into several partitions, each of which may choose between DIP and IBC. We study an intra prediction scheme with the proposed combination, especially the rate-distortion optimization and entropy coding in the scheme. Preliminary experimental results show that the proposed combined intra prediction achieves as high as 5.8% bit-rate saving compared to HEVC anchor.

Discrete Cosine Transform (DCT) has been the commonly used transform for a few decades in image/video coding. However, DCT does not work well on the blocks having anisotropic correlations. In this paper, based on the adaptive dictionary, we propose a new online transform scheme using Orthogonal Matching Pursuit (OMP) for High Efficiency Video Coding (HEVC). For a coding block, we construct its dictionary by exploiting non-local correlations from the reconstructed regions. The OMP algorithm is implemented to obtain the sparse transform coefficients. Experimental results show that the BD-rate savings of the proposed scheme for the sequences with strong edges can be up to 19.9%.

In all of the existing block-based image and video coding standards, blocks are processed in the fixed scan order. Then in HEVC intra coding, intra prediction is always based on the top and/or left neighboring reconstructed pixels, which incurs less accurate prediction for blocks where the spatial correlation is not along the topleft-to-bottomright direction. To obtain better intra prediction, we propose to flexibly determine the coding order of blocks in HEVC intra coding. Complying with the hierarchical quadtree structure in HEVC, our flexible block ordering (FBO) technique recursively decides the coding order of four sub-blocks when splitting one block. Moreover, we propose new methods to perform inter/extrapolation for intra prediction so as to fully utilize neighboring reconstructed pixels, not always being top/left. Experimental results show that our proposed FBO technique achieves on average 2.9% BD-rate reduction compared to HEVC baseline.

Skeleton tracking is a useful and popular application of Kinect. However, it cannot provide accurate reconstructions for complex motions, especially in the presence of occlusion. This paper proposes a new 3-D motion recovery method based on low-rank matrix analysis to correct invalid or corrupted motions. We address this problem by representing a motion sequence as a matrix, and introducing a convex low-rank matrix recovery model, which fixes erroneous entries and finds the correct low-rank matrix by minimizing nuclear norm and norm of constituent clean motion and error matrices. Experimental results show that our method recovers the corrupted skeleton joints, achieving accurate and smooth reconstructions even for complicated motions.

Coded aperture snapshot spectral imaging (CASSI) has been demonstrated as a feasible solution to recover a 3D hyperspectral image by using a single 2D measurement. In this paper, we propose a new hybrid camera design for CASSI to capture high quality hyperspectral images while maintaining the snapshot advantage. Specifically, we employ a complementary RGB camera in conjunction with the CASSI system. The recorded RGB image can provide reliable spectral clue of the scene. By combining the coded hyperspectral information from the CASSI branch and the uncoded color information from the RGB branch, hyperspectral images can be reconstructed with high fidelity. Furthermore, by conducting demosaicing on the raw RGB image as a preprocessing procedure, even better performance can be achieved. Both theoretical analysis and simulation results show improved accuracy of the proposed method compared to the state-of-the-arts.

This paper presents diagonal motion partitions (DMP) for inter prediction in HEVC. In addition to the square and rectangular partitions, we propose to add diagonal shaped partitions to match different motion parts with oblique boundaries. Considering the overlap of pixels along the partition boundaries, the calculation of sum of absolute differences (SAD) and the motion compensation for the pixels on the boundaries are weighted. Besides, the residues of a diagonal prediction unit (PU) are augmented to form a rectangular one, so as to perform Hadamard transform on the residues. We also revise the advanced motion vector prediction (AMVP) and merge candidates based on the diagonal motion partitions. Experimental results show that on average 0.8%-1.0% BD-rate reduction can be achieved by DMP.

No-reference bitstream-layer video quality assessment is very important and practical for monitoring the perceptual experience of end users and facilitating network maintenance. For pervasive Internet Protocol Television and mobile streaming services, in addition to quality degradation due to lossy compression, the unreliable transmission mechanism (i.e., User Datagram Protocol/IP) often leads to quality degradation due to packet loss. Different technical solutions bring in different types of visual artifacts. In this paper, we proposed a hybrid distortion ranking (HDR)-based bitstream-layer quality assessment model, whose artifact combination framework is based on the ranked linear combination operation. The model can predict the perceived quality of a video with sufficient accuracy when the video is distorted by compression artifacts, slicing artifacts, freezing (with frame skipping) artifacts, or their combinations. The core algorithms of the model were adopted into ITU-T Recommendations, P.1202.1 and P.1202.2. Furthermore, with respect to the three different types of artifacts, we compared the proposed no-reference HDR model with some state-of-the-art full-reference perceptual quality assessment models including Video Quality Model (i.e., ITU-T Rec. J.144), structural similarity (SSIM), multiscale SSIM, visual information fidelity, and the widely used metric, peak signal-to-noise ratio. We also compared our HDR model with the top performing no-reference models including Blind/Referenceless Image Spatial Quality Evaluator and video Blind Prediction of Natural Video Quality. The experiment results demonstrate the efficiency of our HDR model.

In contemporary multiuser multiple-input multipleoutput systems, it is common for the reception devices to have a varying number of antennas. When multicast is performed, the number of concurrent spatial streams is limited by the device with the least number of antennas, which prevents more capable devices from getting higher rates. In this paper, we address the antenna heterogeneity in wireless video multicast by the innovative design of multiple similar description (MSD) video coding and multiplexed space-time block coding (M-STBC). MSD coding generates multiple descriptions of a video and features that any linear combinations of the descriptions are decodable. The descriptions comprising of real numbers are further processed by transform and power allocation steps for efficient transmission in a power-constrained system. M-STBC puts symbols in similar descriptions to corresponding space-time positions and ensures decodability under any antenna settings and channel conditions. As a result, we build up a scalable video multicast system, named AirScale, which allows receivers with a various number of antennas to decode from a single transmission, and the reconstructed video quality improves with the number of equipped antennas. Evaluations on Sora shows that, in a {1, 2, 3, 4} × 4 system, AirScale provides baseline quality for one-antenna receiver and a much higher quality for multiantenna receivers. The gain over SoftCast is up to 3.5, 3.9, and 4.1 dB for two-, three-, and four-antenna receivers, respectively.

The Kinect depth sensor has become a versatile tool in different research fields due to its low cost and ease of use, yet its implementation still remains a black box to a large extent. Moreover, the accuracy and speed of Kinect are limited for certain applications. To boost the performance of Kinect while preserving its advantages, we have proposed a depth acquisition method that uses speckle patterns generated by commodity laser emitters as in Kinect, but the speckle density is modulated to allow higher accuracy depth measurement. In this paper, we present the field-programmable gate array (FPGA) design and implementation for the above Kinectlike depth sensing. To achieve a high-speed performance on a normal FPGA (Altera Stratix III EP3SL150), we elaborate the hardware architecture to realize a full-pipeline framework for the two main modules: 1) pixel-wise phase generation and 2) normalized cross-correlation-based phase unwrapping. Experiments show a reconstruction speed of 572 fps at a resolution of 640 × 480 with improved depth accuracy over Kinect and that an even higher resolution is supported. The previous pattern design and the FPGA design in this paper jointly demonstrate the feasibility of next-generation fast, accurate, and low-cost depth sensors.

We investigate a decentralized patrolling problem for dynamic environments where information is distributed alongside threats. In this problem, agents obtain information at a location, but may suffer attacks from the threat at that location. In a decentralized fashion, each agent patrols in a designated area of the environment and interacts with a limited number of agents. Therefore, the goal of these agents is to coordinate to gather as much information as possible while limiting the damage incurred. Hence, we model this class of problem as a transition-decoupled partially observable Markov decision process with health constraints. Furthermore, we propose scalable decentralized online algorithms based on Monte Carlo tree search and a factored belief vector. We empirically evaluate our algorithms on decentralized patrolling problems and benchmark them against the state-of-the-art online planning solver. The results show that our approach outperforms the state-of-the-art by more than 56% for six agents patrolling problems and can scale up to 24 agents in reasonable time.

We propose an analytical model to estimate the depth-error-induced virtual view synthesis distortion (VVSD) in 3D video, taking the distance between reference and virtual views (virtual view position) into account. In particular, we start with a comprehensive preanalysis and discussion over several possible VVSD scenarios. Taking intrinsic characteristic of each scenario into consideration, we specifically classify them into four clusters: 1) overlapping region; 2) disocclusion and boundary region; 3) edge region; and 4) infrequent region. We propose to model VVSD as the linear combination of the distortion under different scenarios (DDSs) weighted by the probability under different scenarios (PDSs). We show analytically that DDS and PDS can be related to the virtual view position using quadratic/biquadratic models and linear models, respectively. Experimental results verify that the proposed model is capable of estimating the relationship between VVSD and the distance between reference and virtual views. Therefore, our model can be used to inform a reference view setup for capturing, or distortion at certain virtual view positions, when depth information is compressed.

The explosion of digital photos has posed a significant challenge to photo storage and transmission for both personal devices and cloud platforms. In this paper, we propose a novel lossless compression method to further reduce the size of a set of JPEG coded correlated images without any loss of information. The proposed method jointly removes inter/intra image redundancy in the feature, spatial, and frequency domains. For each collection, we first organize the images into a pseudo video by minimizing the global prediction cost in the feature domain. We then present a hybrid disparity compensation method to better exploit both the global and local correlations among the images in the spatial domain. Furthermore, the redundancy between each compensated signal and the corresponding target image is adaptively reduced in the frequency domain. Experimental results demonstrate the effectiveness of the proposed lossless compression method. Compared with the JPEG coded image collections, our method achieves average bit savings of more than 31%.

This paper proposes a new image denoising algorithm based on adaptive signal modeling and regularization. It improves the quality of images by regularizing each image patch using bandwise distribution modeling in transform domain. Instead of using a global model for all the patches in an image, it employs content-dependent adaptive models to address the non-stationarity of image signals and also the diversity among different transform bands. The distribution model is adaptively estimated for each patch individually. It varies from one patch location to another and also varies for different bands. In particular, we consider the estimated distribution to have non-zero expectation. To estimate the expectation and variance parameters for every band of a particular patch, we exploit the nonlocal correlation in image to collect a set of highly similar patches as the data samples to form the distribution. Irrelevant patches are excluded so that such adaptively learned model is more accurate than a global one. The image is ultimately restored via bandwise adaptive soft-thresholding, based on a Laplacian approximation of the distribution of similar-patch group transform coefficients. Experimental results demonstrate that the proposed scheme outperforms several state-of-the-art denoising methods in both the objective and the perceptual qualities.

An uncoded transmission scheme called SoftCast has recently shown great potential for wireless video transmission. Unlike conventional approaches, SoftCast processes input images only by a series of transformations and modulates the coefficients directly to a dense constellation for transmission. The transmission is uncoded and lossy in nature, with its noise level commensurate with the channel condition. This paper presents a theoretical analysis for an uncoded visual communication, focusing on developing a quantitative measurements for the efficiency of decorrelation transform in a generalized uncoded transmission framework. Our analysis reveals that the energy distribution among signal elements is critical for the efficiency of uncoded transmission. A decorrelation transform can potentially bring a significant performance gain by boosting the energy diversity in signal representation. Numerical results on Markov random process and real image and video signals are reported to evaluate the performance gain of using different transforms in uncoded transmission. The analysis presented in this paper is verified by simulated SoftCast transmissions. This provide guidelines for designing efficient uncoded video transmission schemes.

The subpixel rendering technology increases the apparent resolution of an LCD/OLED screen by exploiting the physical property that a pixel is composed of RGB individually addressable subpixels. Due to the intrinsic intercoordination between apparent luminance resolution and color fringing artifact, a common method of subpixel image assessment is subjective evaluation. In this paper, we propose a unified subpixel image quality assessment metric called subpixel image assessment (SPA), which syncretizes local subpixel and global pixel features. Specifically, comprehensive subjective studies are conducted to acquire data of user preferences. Accordingly, a collection of low-level features is designed under extensive perceptual validation, capturing subpixel and pixel features, which reflect local details and global distance from the original image. With the features and their measurements as the basis, the SPA is obtained, which leads to a good representation of the subpixel image characteristics. The experimental results justify the effectiveness and the superiority of the SPA. The SPA is also successfully adopted in a variety of applications, including content adaptive sampling and metric-guided image compression.

This research proposes a novel data assisted image transmission scheme, which utilizes a large amount of correlated images stored in the cloud to improve the spectrum efficiency and visual quality. First, a two-layer Coset coding is proposed for the DCT coefficients transmission. The most significant bits (MSB) of the coefficients are generated by the first layer Coset and together with a few low frequency coefficients are transmitted through the most reliable channel coding and digital modulation. The middle bits generated by the second layer Coset are discarded by the sender and the residual bits are transmitted through amplitude modulation. Based on the MSB and the residual bits, an approximation of the original image is reconstructed. With this approximation, a lot of correlated images can be retrieved from the cloud, which are used to recover the discarded middle bits. The two layer Coset coding can significantly decrease the data energy so as to improve the transmission power efficiency. Hence, the end to end distortion of amplitude modulation can be reduced. Second, the image quality can be further improved by joint internal and external denoising with the retrieved images. Simulations show that the proposed scheme outperforms conventional digital schemes about 4 dB in peak signal to noise power ratio (PSNR) and achieves 2 dB gain over the state-of-the-art uncoded transmission. At low signal to noise power ratio (SNR), an additional 2-3 dB gain is achieved. The visual quality comparison also validates the objective image assessment result.

We propose a novel dual-camera design to acquire 4D high-speed hyperspectral (HSHS) videos with high spatial and spectral resolution. Our work has two key technical contributions. First, we build a dual-camera system that simultaneously captures a panchromatic video at a high frame rate and a hyperspectral video at a low frame rate, which jointly provide reliable projections for the underlying HSHS video. Second, we exploit the panchromatic video to learn an over-complete 3D dictionary to represent each band-wise video sparsely, and a robust computational reconstruction is then employed to recover the HSHS video based on the joint videos and the self-learned dictionary. Experimental results demonstrate that, for the first time to our knowledge, the hyperspectral video frame rate reaches up to 100fps with decent quality, even when the incident light is not strong.

We present a compound image compression scheme based on the dictionary-based Lempel-Ziv-Markov chain algorithm (LZMA), under the framework of High Efficiency Video Coding (HEVC). Through matching strings from the sliding window dictionary, LZMA exploits the characteristics of the repeated patterns over the text and graphics regions of compound images, and represents them compactly. To obtain high compression efficiency even for noisy text and graphics contents, we have modified LZMA to support both lossless and lossy compression. We develop and treat it as a new intramode of HEVC. Experimental results show that the proposed scheme achieves significant coding gains for compound image compression. Thanks to the introduction of the lossy LZMA, the compression performance for noisy compound images is improved for more than 5dB in terms of PSNR in comparison with the lossless LZMA scheme.

Redundancy is necessary for a storage system to recover from errors. The frequent errors in large-scale systems, e.g. cloud, make it desired to reduce the recovery cost. Among all kinds of data stored in the cloud, video takes a large portion due to its large data volume. The other characteristic of video is that a certain distortion can be tolerated. This paper investigates using scalable video representation and unequal error protection scheme to reduce the storage and recovery costs in the cloud. By introducing more protection for the base layer and less on the enhancement layers, it can achieve a better tradeoff between storage and reconstruction costs although the reliability for the enhancement layer sacrifices a little. Simulation results based on local reconstruction codes (LRC) show that comparing with the existing (12, 2, 2) LRC code in Windows Azure Storage, the reconstruction cost can be reduced from 6x to 3x at the same storage cost at the expense of possible video quality loss.

Depth sensors based on Time-of-Flight (ToF) and Phase Shifting (PS) have complementary strengths and weaknesses. ToF can provide real-time depth but limited in resolution and sensitive to noise. PS can generate accurate and robust depth with high resolution but requires a number of patterns that leads to high latency. In this paper, we propose a novel fusion framework to take advantages of both ToF and PS. The basic idea is using the coarse depth from ToF to disambiguate the wrapped depth from PS. Specifically, we address two key technical problems: cross-modal calibration and interference-free synchronization between ToF and PS sensors. Experiments demonstrate that the proposed method generates accurate and robust depth with high resolution and low latency, which is beneficial to tremendous applications.

Efficient and robust wireless stereo video delivery is an enabling technology for various mobile 3D applications. Existing digital solutions have high source coding efficiency but are not robust to channel variations, while analog solutions have the opposite characteristics. In this paper, we design a novel hybrid digital-analog (HDA) solution to embrace the advantages of both solutions and avoid their drawbacks. Basically, in each pair of stereo frames, one frame is digitally encoded to ensure basic quality and the other is analogly processed to opportunistically utilize good channels for better quality. To improve the system efficiency, we design a zigzag coding structure such that both intra-view and inter-view correlations can be explored through prediction in the frames to be analogly coded. A reference selection mechanism is proposed to further improve the coding efficiency. In addition, we address the problem of optimal power and bandwidth allocation between digital and analog streams. We implement a system, named Swift, and perform extensive trace-driven evaluations based on a software-defined radio platform. We show that Swift outperforms an omniscient digital scheme under the same bandwidth and power constraints, or can have around 2x power saving in order to achieve comparable performance. Subjective quality assessment evidences that Swift provides significantly better visual quality than a straightforward HDA extension of SoftCast.

In this paper, we present the system design and the key techniques of our mobile robot platform called KeJia-LC , who won the first place in the demo challenge on Benchmarkinng Service Robots in RoboCup 2015. Given the fact that KeJia-LC is a low-cost version of our KeJia robot without shoulder and arm, several new technical demands comparing to RoboCup@Home are highlighted for better understanding of our system. With the elaborate design of hardware and the reasonable selection of sensors, our robot platform has the features of low cost, wide generality and good extensibility. Moreover, we integrate several functional softwares (such as 2D&3D mapping, localization and navigation) following the competition rules, which are critical to the performance of our robot. The effectiveness and robustness of our robot system has been proven in the competition.

Most of existing benchmarking tools for service robots are basically qualitative, in which a robot’s performance on a task is evaluated based on completion/incompletion of actions contained in the task. In the effort reported in this paper, we tried to implement a synthetical benchmarking system on domestic mobile platforms. Synthetical benchmarking consists of both qualitative and quantitative aspects, such as task completion, accuracy of task completions and efficiency of task completions, about performance of a robot. The system includes a set of algorithms for collecting, recording and analyzing measurement data from a MoCap system. It was used as the evaluator in a competition called the BSR challenge, in which 10 teams participated, at RoboCup 2015. The paper presents our motivations behind synthetical benchmarking, the design considerations on the synthetical benchmarking system, the realization of the competition as a comparative study on performance evaluation of domestic mobile platforms, and an analysis of the teams’ performance. These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Mobile video live streaming is facing great challenges in offering high quality of experience (QoE) under varying channel conditions. In this paper, we propose a progressive pseudo-analog transmission scheme in which the received video quality gracefully adapts to both SNR and bandwidth variations. Building upon the emerging pseudoanalog video transmission, the proposed scheme further adopts a greedy approach to improve the received video quality with each allocated bandwidth share. The optimal scheduling and power allocation are derived under the mean squared error (MSE) criterion. Testbed evaluations show that the proposed scheme outperforms the state-of-the-art digital and analog transmission schemes by a notable margin.

This paper presents an effective method to enhance the quality of dim light surveillance via gradient fusion. We simply take the advantage that surveillance cameras capture a large quantity of valuable information at the same viewpoint during the day. And it can be used to make the video at night easier to perceive. Based on a gradient domain technique, all the important local perceptual cues from the original video are automatically combined with the supporting daytime context, while avoiding traditional problems such as aliasing, ghosting and haloing. Experimental results show that our method outperforms the state-of-the-art ones, and can even handle some challenging conditions without altering the parameters.

This paper proposes a distributed compressive sensing (CS) scheme for robust image transmission over unknown or time-varying channels with highly correlated images at the decoder. A compressed thumbnail is first transmitted after digital forward error correction (FEC) and modulation to retrieve highly correlated images and generate a side information (SI) at the decoder. The current residual image after subtracting the decompressed thumbnail is then coded and transmitted by CS through a very dense constellation without FEC. The linear representation of the residual signal by CS measurements and rateless sampling makes it able to achieve graceful degradation and bandwidth scalability without channel feedback. Moreover, a transform-domain power allocation is employed before random sampling to protect against channel errors. At the decoder, both the nonlocal correlations within the original image and the correlation with the SI are exploited in CS decoding via a low-rank regulation on similar patches. After CS decoding, a block-wise minimum-mean-square-error (MMSE) reconstruction using the SI is further performed in the spatial domain to enhance the reconstruction quality. Simulations on landmark images and an unknown Gaussian channel show that an up to 10 dB gain is achieved at low channel SNRs compared with the state-of-the-art uncoded image transmission scheme, i.e. SoftCast, when highly correlated images are available at the decoder.

The key problem in photo album compression is how to exploit the correlation among the images. In this paper, we propose a novel incremental structure from motion (SfM) based prediction method for lossless photo album compression. Unlike the previous methods, we exploit the redundancy among images through their inherent geometric relationship generated by SfM. Based on the point cloud and camera poses, each prediction image is generated by projecting, triangulation and warping. Finally, the target image is compressed by an HEVC-like encoder with the prediction image as main reference. Experimental results demonstrate the advantage of our method, especially for images with scenes containing complicated geometric structures.

Screen content videos often have mixed content consisting of various types such as natural content, text and graphics in the same picture. To achieve high coding efficiency for this content, new coding tools are developed in the High Efficiency Video Coding (HEVC) standard Screen Content Coding (SCC) extension. Among them, intra block copy (IntraBC) allows a nonlocal intra prediction from the coded region of the same picture. However, the rate-distortion optimization (RDO) scheme of the screen content coding still follows that of the HEVC reference software and each block is optimized locally. The screen content characteristics are not fully utilized in the current RDO scheme. This paper presents a weighted RDO scheme for the intra coding of screen content videos, which measures the importance of each block within the picture first and then larger distortion weights are applied to the blocks with larger importance in the RDO. In this way, a better rate-distortion trade-off can be achieved for the picture instead of the blocks themselves. The experimental results show that up to 5.5% coding efficiency gain can be achieved compared with the reference software.

Wireless video broadcast poses a challenge to the conventional visual communication in providing simultaneously each receiver the best video quality under its channel condition. Soft video broadcast, as a newly emerged wireless video broadcast scheme, is able to accommodate multiple receivers of different channel SNRs. However, the current soft video broadcast frameworks such as SoftCast require the bandwidth of the wireless channel to match the number of video coefficients per second. When the channel bandwidth is larger, the existing frameworks become not very efficient in bandwidth expansion. More importantly, it is possible that the users in broadcast applications have different bandwidths. However, none of the existing soft video broadcast frameworks considers bandwidth heterogeneity. In this paper, we propose a soft video broadcast framework, called LayerCast, which can simultaneously accommodate heterogeneous users with diverse SNRs and diverse bandwidths. The bandwidth expansion problem is solved by applying layered coset coding. More importantly, we derive a globally optimal power allocation between layers and, within each layer, between each DCT chunk. In simulations, the proposed framework outperforms SoftCast of up to 4 dB in video PSNR, and outperforms H.264-based framework up to 8 dB in broadcast.

With multimedia flourishing on the Web, it is easy to find similar images for a query, especially landmark images. Traditional image coding, such as JPEG, cannot exploit correlations with external images. Existing vision-based approaches are able to exploit such correlations by reconstructing from local descriptors but cannot ensure the pixel-level fidelity of the reconstruction. In this paper, a cloud-based distributed image coding (Cloud-DIC) scheme is proposed to exploit external correlations for mobile photo uploading. For each input image, a thumbnail is transmitted to retrieve correlated images and reconstruct it in the cloud by geometrical and illumination registrations. Such a reconstruction serves as the side information (SI) in the Cloud-DIC. The image is then compressed by a transform-domain syndrome coding to correct the disparity between the original image and the SI. Once a bitplane is received in the cloud, an iterative refinement process is performed between the final reconstruction and the SI. Moreover, a joint encoder/decoder mode decision at block, frequency, and bitplane levels is proposed to adapt to different correlations. Experimental results on a landmark image database show that the Cloud-DIC can largely enhance the coding efficiency both subjectively and objectively, with up to 5-dB gains and 70% bits saving over JPEG with arithmetic coding, and perform comparably at low bitrates with the intra coding of the High Efficiency Video Coding standard with a much lower encoder complexity.

Although the High Efficiency Video Coding (HEVC) standard significantly improves the coding efficiency of video compression, it is unacceptable even in offline applications to spend several hours compressing 10 s of high-definition video. In this paper, we propose using a multicore central processing unit (CPU) and an off-the-shelf graphics processing unit (GPU) with 3072 streaming processors (SPs) for HEVC fast encoding, so that the speed optimization does not result in loss of coding efficiency. There are two key technical contributions in this paper. First, we propose an algorithm that is both parallel and fast for the GPU, which can utilize 3072 SPs in parallel to estimate the motion vector (MV) of every prediction unit (PU) in every combination of the coding unit (CU) and PU partitions. Furthermore, the proposed GPU algorithm can avoid coding efficiency loss caused by the lack of a MV predictor (MVP). Second, we propose a fast algorithm for the CPU, which can fully utilize the results from the GPU to significantly reduce the number of possible CU and PU partitions without any coding efficiency loss. Our experimental results show that compared with the reference software, we can encode high-resolution video that consumes 1.9% of the CPU time and 1.0% of the GPU time, with only a 1.4% rate increase.

Deblurring saturated night images are a challenging problem because such images have low contrast combined with heavy noise and saturated regions. Unlike the deblurring schemes that discard saturated regions when estimating blur kernels, this paper proposes a novel scheme to deduce blur kernels from saturated regions via a novel kernel representation and advanced algorithms. Our key technical contribution is the proposed function-form representation of blur kernels, which regularizes existing matrix-form kernels using three functional components: 1) trajectory; 2) intensity; and 3) expansion. From automatically detected saturated regions, their skeleton, brightness, and width are fitted into the corresponding three functional components of blur kernels. Such regularization significantly improves the quality of kernels deduced from saturated regions. Second, we propose an energy minimizing algorithm to select and assign the deduced function-form kernels to partitioned image regions as the initialization for non-uniform deblurring. Finally, we convert the assigned function-form kernels into matrix form for more detailed estimation in a multi-scale deconvolution. Experimental results show that our scheme outperforms existing schemes on challenging real examples.

Single image denoising suffers from limited data collection within a noisy image. In this paper, we propose a novel image denoising scheme, which explores both internal and external correlations with the help of web images. For each noisy patch, we build internal and external data cubes by finding similar patches from the noisy and web images, respectively. We then propose reducing noise by a two-stage strategy using different filtering approaches. In the first stage, since the noisy patch may lead to inaccurate patch selection, we propose a graph based optimization method to improve patch matching accuracy in external denoising. The internal denoising is frequency truncation on internal cubes. By combining the internal and external denoising patches, we obtain a preliminary denoising result. In the second stage, we propose reducing noise by filtering of external and internal cubes, respectively, on transform domain. In this stage, the preliminary denoising result not only enhances the patch matching accuracy but also provides reliable estimates of filtering parameters. The final denoising image is obtained by fusing the external and internal filtering results. Experimental results show that our method constantly outperforms state-of-the-art denoising schemes in both subjective and objective quality measurements, e.g., it achieves >2 dB gain compared with BM3D at a wide range of noise levels.

Markov decision processes (MDPs) provide a rich framework for planning under uncertainty. However, exactly solving a large MDP is usually intractable due to the “curse of dimensionality”— the state space grows exponentially with the number of state variables. Online algorithms tackle this problem by avoiding computing a policy for the entire state space. On the other hand, since online algorithm has to find a near-optimal action online in almost real time, the computation time is often very limited. In the context of reinforcement learning, MAXQ is a value function decomposition method that exploits the underlying structure of the original MDP and decomposes it into a combination of smaller subproblems arranged over a task hierarchy. In this article, we present MAXQ-OP—a novel online planning algorithm for large MDPs that utilizes MAXQ hierarchical decomposition in online settings. Compared to traditional online planning algorithms, MAXQ-OP is able to reach much more deeper states in the search tree with relatively less computation time by exploiting MAXQ hierarchical decomposition online. We empirically evaluate our algorithm in the standard Taxi domain—a common benchmark for MDPs—to show the effectiveness of our approach. We have also conducted a long-term case study in a highly complex simulated soccer domain and developed a team named WrightEagle that has won five world champions and five runners-up in the recent 10 years of RoboCup Soccer Simulation 2D annual competitions. The results in the RoboCup domain confirm the scalability of MAXQ-OP to very large domains.

Hybrid digital-analog (HDA) transmission has gained increasing attention recently in the context of wireless video delivery , for its ability to simultaneously achieve high transmission efficiency and smooth quality adaptation. However, previous systems are optimized solely based on the mean squared error criterion without taking the perceptual video quality into consideration. In this work, we propose a structure-preserving HDA video delivery system, named SharpCast, to improve both the objective and subjective visual quality. SharpCast decomposes a video into a content part and structure part. The latter is important to the human perception and therefore is protected with a robust digital transmission scheme. Then, the energy-intensive part in the content information is extracted and transmitted in digital for energy efficiency while the residual is transmitted in analog to achieve the desired smooth adaptation. We formulate the resource (power and bandwidth) allocation problem in SharpCast and solve the problem with a greedy strategy. Evaluations over nine standard 720p video sequences show that the proposed SharpCast system outperforms the state-of-the-art digital, analog, and HDA schemes by a notable margin in both peak signal-to-noise ratio (PSNR) and structural similarity (SSIM).

This paper addresses the data persistence problem in wireless sensor networks (WSNs) where static sinks are not present and the sensed data have to be temporarily but resiliently stored in the network. Based on the observation that sensor readings are correlated, we propose compressive data persistence (CDP) scheme that makes use of the compressive sensing (CS) theory. Each sensor node independently computes and stores a random projection of the sensed data, such that a mobile sink can recover the data with high probability after visiting a small and random portion of the network. As a prerequisite of distributed CS encoding, sensor readings from all nodes are disseminated within the network through random walk. Therefore, the CS measurement matrix depends heavily on how the random walk is performed. In this paper, we present an in-depth analysis on the interplay between random walk parameters and sensing data characteristics, and derive the conditions in successful CS data recovery. In addition, we discover that there is a trade-off between the number of random walk instances and steps in order to achieve the required data persistence performance. Experiments using real sensor data verify that the proposed CDP scheme achieves much lower decoding ratio than the state-of-the-art Fountain code based schemes or the decentralized erasure codes based schemes, and demonstrate that there exist energy-optimized random walk parameters for CDP.

Many multi-agent coordination problems can be represented as DCOPs. Motivated by task allocation in disaster response, we extend standard DCOP models to consider uncertain task rewards where the outcome of completing a task depends on its current state, which is randomly drawn from unknown distributions. The goal of solving this problem is to find a solution for all agents that minimizes the overall worst-case loss. This is a challenging problem for centralized algorithms because the search space grows exponentially with the number of agents and is nontrivial for existing algorithms for standard DCOPs. To address this, we propose a novel decentralized algorithm that incorporates Max-Sum with iterative constraint generation to solve the problem by passing messages among agents. By so doing, our approach scales well and can solve instances of the task allocation problem with hundreds of agents and tasks.

In this paper, we deal with the image deblurring problem in a completely new perspective by proposing separable kernel to represent the inherent properties of the camera and scene system. Specifically, we decompose a blur kernel into three individual descriptors (trajectory, intensity and point spread function) so that they can be optimized separately. To demonstrate the advantages, we extract one-pixel-width trajectories of blur kernels and propose a random perturbation algorithm to optimize them but still keeping their continuity. For many cases, where current deblurring approaches fall into local minimum, excellent deblurred results and correct blur kernels can be obtained by individually optimizing the kernel trajectories. Our work strongly suggests that more constraints and priors should be introduced to blur kernels in solving the deblurring problem because blur kernels have lower dimensions than images.

In this paper, we propose a novel two-step scheme to filter heavy noise from images with the assistance of retrieved Web images. There are two key technical contributions in our scheme. First, for every noisy image block, we build two three dimensional (3D) data cubes by using similar blocks in retrieved Web images and similar nonlocal blocks within the noisy image, respectively. To better use their correlations, we propose different denoising strategies. The denoising in the 3D cube built upon the retrieved images is performed as median filtering in the spatial domain, whereas the denoising in the other 3D cube is performed in the frequency domain. These two denoising results are then combined in the frequency domain to produce a denoising image. Second, to handle heavy noise, we further propose using the denoising image to improve image registration of the retrieved Web images, 3D cube building, and the estimation of filtering parameters in the frequency domain. Afterwards, the proposed denoising is performed on the noisy image again to generate the final denoising result. Our experimental results show that when the noise is high, the proposed scheme is better than BM3D by more than 2 dB in PSNR and the visual quality improvement is clear to see.

Conventional image and video communication systems are usually designed with the objective being to maximize the fidelity of reconstructed images measured by mean square errors (MSE). It is well known that the fidelity metric MSE may not reflect the visual quality perceived by human eyes. Recent advancements in image quality assessment tell us that the structural similarity (SSIM), especially the gradient similarity, reveals the perceptual fidelity of images more reliably. Inspired by this observation, this paper proposes a new image communication approach, which conveys the visual information in an image by transmitting the image gradients and recovers the image from the received gradient data at decoder side using statistical image prior knowledge. In particular, we designed a gradient-based image SoftCast scheme for wireless scenarios. Experimental results show that the proposed scheme can produce reconstruction images with much better perceptual quality. The advantage in perceptual quality is verified by the quality improvement measured by the metrics SSIM and gradient signal-to-noise ratio (GSNR).

This paper proposes a cloud-based distributed image coding scheme (Cloud-DIC) to exploit the strong correlations with external partial-duplicate images in the cloud. It features both high coding efficiency and low encoder complexity, which makes it suitable for photo sharing on mobile devices. To get the side information in the cloud, a thumbnail of the current image is transmitted to retrieve highly correlated images and reconstruct through geometrical registration and adaptive patched-based stitching. The current image is then compressed by a transform-domain syndrome coding, bitplane by bitplane. Once a bitplane is received, the decoded high-quality image is further used to refine the side information in the cloud, which will benefit the coding of following bitplanes and the reconstruction. Experimental results on a landmark image database show that it can largely enhance the coding efficiency both subjectively and objectively with up to 5 dB gains and 58% bits saving over JPEG.

For wireless scenarios where the channel condition fluctuates unpredictably, a novel image/video communication scheme, named SoftCast, was recently proposed to provide graceful quality degradation and competitive performance simultaneously. Unlike conventional approaches, SoftCast decorrelates input images by a transform and modulates the coefficients directly to a dense constellation for transmission, leaving out the conventional quantization, entropy coding and channel coding. The transmission is lossy in nature, with its noise level commensurate with the channel condition. To reconstruct images from the received noisy data, SoftCast employs a linear least-square estimator (LLSE), but it tends to produce annoying reconstruction artifacts. This paper proposes a high-quality image reconstruction algorithm for SoftCast, employing a collaborative estimator to utilize both the local correlation and non-local similarity within images. Experimental results show that the proposed method outperforms the existing SoftCast scheme, achieving remarkable improvement in the objective and subjective qualities of the reconstruction images.

This research studies robust uncoded video transmission over wireless fast fading channel, where only statistical channel state information (CSI) is available at the transmitter. We observe that increasing channel diversity for high priority (HP) data is essential to improving the robustness of video transmission in fading channels. By utilizing the noise and loss resilient nature of video, we find it possible to design a more robust system by re-allocating the power and channel uses among HP and LP (low priority) data. With total power and channel use constraints, we derive an optimal resource allocation scheme under the squared error distortion criterion. In particular, we first propose a new power allocation algorithm at given channel allocation. Second, based on the proposed power allocation algorithm, we design a channel allocation algorithm to strike the tradeoff between the diversity increase of HP data and the information loss of LP data. Third, under known noise power distribution, we derive the optimal resource allocation for uncoded video multicast. Simulations show that the proposed system achieves 2dB and 5dB gain in average and outage PSNR over Softcast in video unicast, and around 1.4dB and 4dB gain in multicast.

This paper introduces 1-D dictionary mode designed for screen content coding. Two 1-D dictionary modes are designed to improve the coding efficiency for screen content. The first one is called normal dictionary mode, in which a virtual dictionary should be maintained and all the prediction comes from the virtual dictionary. The other one is called reconstruction based dictionary mode, where no virtual dictionary is to be maintained and all the previously reconstructed pixels in the same picture can be used for prediction. Hash based search is designed to find matching for both dictionary modes efficiently. 1-D dictionary mode with variable block sizes are also supported in the proposed scheme. The experimental results show the proposed algorithm achieves about 10% ~ 18.4% bit saving for different coding structures. The bit saving is up to 60% for the proposed method.

This paper introduces a unified framework of hash-based matching method for screen content coding. Screen content has some different characteristics from camera-captured content, such as large motion and repeating patterns. Hash-based matching is proposed to better explore the correlation in screen content, thus, improving the coding efficiency. The proposed method can handle both intra picture and inter picture block matching with variable block sizes in a unified framework. The proposed framework is also easy to be extended to handle other motion models to further improve the coding efficiency of screen content. We also develop fast encoding algorithms to make full use of the hash results. The experimental results show the proposed algorithm achieves about 12% bit saving while saving more than 25% encoding time. The bit saving is up to 57% and the encoding time saving is up to 60% for the proposed method.

The explosion in digital photography poses a significant challenge when it comes to photo storage for both personal devices and the Internet. In this paper, we propose a novel lossless compression method to further reduce the storage size of a set of JPEG coded correlated images. In this method, we propose jointly removing the inter-image redundancy in the feature, spatial, and frequency domains. For each album, we first organize the images into a pseudo video by minimizing the global predictive cost in the feature domain. We then introduce a disparity compensation method to enhance the spatial correlation between images. Finally, the redundancy between the compensated signal and the corresponding target image is adaptively reduced in the frequency domain. Moreover, our proposed scheme is able to losslessly recover not only raw images but also JPEG files. Experimental results demonstrate the efficiency of our proposed lossless compression, which achieves more than 12% bit-saving on average compared with JPEG coded albums.

High Efficiency Video Coding (HEVC) uses a very flexible tree structure to organize coding units, which leads to a superior coding efficiency compared with previous video coding standards. However, such a flexible coding unit tree structure also places a great challenge for encoders. In order to fully exploit the coding efficiency brought by this structure, huge amount of computational complexity is needed for an encoder to decide the optimal coding unit tree for each image block. One way to achieve this is to use parallel computing enabled by many-core processors. In this paper, we analyze the challenge to use many-core processors to make coding unit tree decision. Through in-depth understanding of the dependency among different coding units, we propose a parallel framework to decide coding unit trees. Experimental results show that, on the Tile64 platform, our proposed method achieves averagely more than 11 and 16 times speedup for 1920x1080 and 2560x1600 video sequences, respectively, without any coding efficiency degradation.

This research addresses the problem of robust linear video transmission over the Rayleigh fading channel, where only statistical channel state information (CSI) is available to the sender. We observe that discarding low-priority (LP) data and saving the channel uses for high-priority (HP) data can significantly improve the quality of the received video. We formulate an optimization problem that aims to minimize the total squared error of a multi-variant Gaussian random vector under the given bandwidth and power resources. To tame the complexity of this NP-hard problem, we analyze two sub-problems, namely power allocation and bandwidth allocation, and propose an iterative algorithm to approximate the solution. Subsequently, we propose a one-pass two-step fast algorithm that further reduces both algorithmic and computational complexity. A linear video transmission system is implemented based on the proposed algorithm. Simulations show that our system significantly outperforms Soft-Cast, and the PSNR gain at 5th percentile of 1000 test runs is between 4.0 dB and 7.5 dB under varying noise levels.

High Efficiency Video Coding (HEVC) provides superior coding efficiency than previous video coding standards at the cost of increasing encoding complexity. The complexity increase of motion estimation (ME) procedure is rather significant, especially when considering the complicated partitioning structure of HEVC. To fully exploit the coding efficiency brought by HEVC requires a huge amount of computations. In this paper, we analyze the ME structure in HEVC and propose a parallel framework to decouple ME for different partitions on many-core processors. Based on local parallel method (LPM), we first use the directed acyclic graph (DAG)-based order to parallelize coding tree units (CTUs) and adopt improved LPM (ILPM) within each CTU (DAGILPM), which exploits the CTU-level and prediction unit (PU)-level parallelism. Then, we find that there exist completely independent PUs (CIPUs) and partially independent PUs (PIPUs). When the degree of parallelism (DP) is smaller than the maximum DP of DAGILPM, we process the CIPUs and PIPUs, which further increases the DP. The data dependencies and coding efficiency stay the same as LPM. Experiments show that on a 64-core system, compared with serial execution, our proposed scheme achieves more than 30 and 40 times speedup for 1920 × 1080 and 2560 × 1600 video sequences, respectively.

In this paper, we propose a novel coding and transmission scheme, called LineCast, for broadcasting satellite images to a large number of receivers. The proposed LineCast matches perfectly with the line scanning cameras that are widely adopted in orbit satellites to capture high-resolution images. On the sender side, each captured line is immediately compressed by a transform-domain scalar modulo quantization. Without syndrome coding, the transmission power is directly allocated to quantized coefficients by scaling the coefficients according to their distributions. Finally, the scaled coefficients are transmitted over a dense constellation. This line-based distributed scheme features low delay, low memory cost, and low complexity. On the receiver side, our proposed line-based prediction is used to generate side information from previously decoded lines, which fully utilizes the correlation among lines. The quantized coefficients are decoded by the linear least square estimator from the received data. The image line is then reconstructed by the scalar modulo dequantization using the generated side information. Since there is neither syndrome coding nor channel coding, the proposed LineCast can make a large number of receivers reach the qualities matching their channel conditions. Our theoretical analysis shows that the proposed LineCast can achieve Shannon's optimum performance by using a high-dimensional modulo-lattice quantization. Experiments on satellite images demonstrate that it achieves up to 1.9-dB gain over the state-of-the-art 2D broadcasting scheme and a gain of more than 5 dB over JPEG 2000 with forward error correction.

Time multiplexing (TM) and spatial neighborhood (SN) are two mainstream structured light techniques widely used for depth sensing. The former is well known for its high accuracy and the latter for its low delay. In this paper, we explore a new paradigm of scalable depth sensing to integrate the advantages of both the TM and SN methods. Our contribution is twofold. First, we design a set of hybrid structured light patterns composed of phase-shifted fringe and pseudo-random speckle. Under the illumination of the hybrid patterns, depth can be decently reconstructed either from a few consecutive frames with the TM principle for static scenes or from a single frame with the SN principle for dynamic scenes. Second, we propose a scene-adaptive depth sensing framework based on which a global or region-wise optimal depth map can be generated through motion detection. To validate the proposed scalable paradigm, we develop a real-time (20 fps) depth sensing system. Experimental results demonstrate that our method achieves an efficient balance between accuracy and speed during depth sensing that has rarely been exploited before.

We have observed two trends, growing wireless capability at the physical layer powered by MIMO-OFDM and growing video traffic as the dominant application traffic. Both the video source and MIMO-OFDM channel components exhibit nonuniform energy distribution. This has motivated us to leverage the source data redundancy at the channel to achieve high video recovery performance. We propose ParCast+ that first separates the source and the channel into independent components, matches the more important source components with higher-gain channel components, allocates power weights with joint consideration to the source and the channel, and uses pseudo-analog modulation for transmission. Such a scheme achieves fine-grained unequal error protection across source components. We implemented ParCast+ in Matlab and on Sora. Extensive evaluation has shown that our scheme outperforms competing schemes by notable margins, sometimes up to 6.4 dB in PSNR for challenging scenarios.

3D model-based object recognition has been a noticeable research trend in recent years. Common methods find 2D-to-3D correspondences and make recognition decisions by pose estimation, whose efficiency usually suffers from noisy correspondences caused by the increasing number of target objects. To overcome this scalability bottleneck, we propose an efficient 2D-to-3D correspondence filtering approach, which combines a light-weight neighborhood-based step with a finer-grained pairwise step to remove spurious correspondences based on 2D/3D geometric cues. On a dataset of 300 3D objects, our solution achieves ~10 times speed improvement over the baseline, with a comparable recognition accuracy. A parallel implementation on a quad-core CPU can run at ~3fps for 1280×720 images.

This paper proposes novel density modulated binary patterns for depth acquisition. Similar to Kinect, the illumination patterns do not need a projector for generation and can be emitted by infrared lasers and diffraction gratings. Our key idea is to use the density of light spots in the patterns to carry phase information. Two technical problems are addressed here. First, we propose an algorithm to design the patterns to carry more phase information without compromising the depth reconstruction from a single captured image as with Kinect. Second, since the carried phase is not strictly sinusoidal, the depth reconstructed from the phase contains a systematic error. We further propose a pixel-based phase matching algorithm to reduce the error. Experimental results show that the depth quality can be greatly improved using the phase carried by the density of light spots. Furthermore, our scheme can achieve 20 fps depth reconstruction with GPU assistance.

In this paper, we present a novel single-shot method for dense and accurate 3D scanning. Our method takes advantage of two conventional techniques, i.e., stereo and Fourier fringe analysis (FFA). While FFA is competent for high-density and high-precision phase measurement, stereo solves the phase ambiguity caused by the periodicity of the fringe. By jointly using the intensity images and unwrapped phase maps from stereo, the pixel-wise absolute depth can be obtained through a sparse matching process efficiently and reliably. Due to its single-shot property and low complexity, the proposed method facilitates dense and accurate 3D scanning in time-critical applications.

Bag-of-visual words (BoW) representation has been widely used in the large scale image retrieval. Though efficient, it ignores the geometric correlation among visual words, whereas the geometric verification has demonstrated its effectiveness in image retrieval. In this paper, we propose a new representation - visual group, to improve the retrieval precision by grouping the geometrically related features based on the inclusion relationship between features at different scales. A visual group consists of a master feature and several member features covered by the master feature. The geometric constraint inside each group is introduced into visual group matching for efficient geometric verification. Experimental evaluation on the dataset Oxford5K+Flickr1M shows that our visual group based image search approach outperforms BoW and the state-of-the-art visual phrase based schemes.

The biggest challenge in image set compression is how to efficiently remove the set redundancy among images as well as the redundancy inside a single image. Different from all the previous schemes, in this paper we are the first to propose a generic image set compression scheme which removes the set redundancy based on local features in addition to luminance values. The SIFT (Scale Invariant Feature Transform) descriptor which characterizes an image region invariant to scale and rotation is utilized in our scheme to measure and further enhance the correlation among images. Given an image set, we build a minimal cost prediction structure according to the SIFT-based prediction measure between images. We also utilize a SIFT-based global transformation to enhance the correlation between two images by aligning them to each other in terms of both geometry and intensity. The set redundancy and image redundancy are both further reduced by block-based motion estimation and rate-distortion optimal mechanism proposed in HEVC. Experimental results show that our new feature based scheme always produces the best result regardless the image set's properties.

In wireless scenarios where the channel condition may vary drastically, visual communication systems using source and channel coding generally suffer from threshold effect. An uncoded transmission scheme called SoftCast [1]-[3], however, was recently shown to provide both graceful quality transition and competitive performance. In SoftCast, image signal is directly modulated to a dense constellation using proper power for transmission, solely after employing a transform for energy compaction, leaving out conventional quantization, entropy coding and channel coding. The received signal is lossy in nature, with its noise level commensurate with the channel condition. This paper presents a theoretical analysis for uncoded visual communication, focusing on the role of transform and the quantitative measurement of transform gain in a generalized uncoded transmission framework with optimal power allocation. Our analysis reveal that the energy distribution among signal elements plays an important role in the power-distortion performance. Further analysis show that the energy compaction capability of decorrelation transform can bring significant gain by boosting the energy diversity in signal representation. Numerical analysis results are reported for Markov random signals and natural images, respectively. The performance of typical transforms, e.g. KLT, DCT and DWT, and the effect of different transform sizes or levels are evaluated. These analysis results are verified by simulations.

This paper presents a new exemplar-based image super-resolution (SR) method in which we propose making use of scale invariant image features for high frequency (HF) approximation. We introduce the scale invariant feature transform (SIFT) descriptors in both building an exemplar dataset adaptively and producing the HF details with respect to the features of an input low resolution image. Given a large image database, we propose using the highly correlated images retrieved by SIFT descriptors for exemplar training rather than using a general set of images to increase the matching accuracy. Through building the training set of high resolution/low resolution exemplar pairs, the HF details for SR are retrieved from the training set by matching the SIFT features in a dense way. The flexibility as well as effectiveness of our SR approach is demonstrated at different magnification factors, e.g. 3 and 4. Experimental results show that our SIFT-based SR approach achieves enhanced high resolution images in terms of both objective and subjective qualities in comparison with the state-of-the-art exemplar-based methods.

This paper challenges the conventional wisdom that video redundancy should be removed as much as possible for efficient communications. We discover that, by keeping spatial redundancy at the sender and properly utilizing it at the receiver, we can build a more robust and even more efficient wireless video communication system than existing ones.

Monte-Carlo tree search is drawing great interest in the domain of planning under uncertainty, particularly when little or no domain knowledge is available. One of the central problems is the trade-off between exploration and exploitation. In this paper we present a novel Bayesian mixture modelling and inference based Thompson sampling approach to addressing this dilemma. The proposed Dirichlet-NormalGamma MCTS (DNG-MCTS) algorithm represents the uncertainty of the accumulated reward for actions in the MCTS search tree as a mixture of Normal distributions and inferences on it in Bayesian settings by choosing conjugate priors in the form of combinations of Dirichlet and NormalGamma distributions. Thompson sampling is used to select the best action at each decision node. Experimental results show that our proposed algorithm has achieved the state-of-the-art comparing with popular UCT algorithm in the context of online planning for general Markov decision processes.

Phase shifting is a widely used method for accurate and dense 3D reconstruction. However, at least three images of the same scene are required for each reconstruction, so measurement errors are inevitable in dynamic scenes, even with high-speed hardware. In this paper, we propose a Fourier transform assisted phase shifting method to overcome the motion vulnerability in phase shifting. A new model with motion-related phase shifts is formulated, and the coarse phase measurements obtained by Fourier transform profilemetry are used to estimate the unknown phase shifts. The phase errors caused by motion are greatly reduced in this way. Experimental results show that the proposed method can obtain accurate and dense 3D reconstruction of dynamic scenes, with regard to different kinds of motion.

The key task in image set compression is how to efficiently remove set redundancy among images and within a single image. In this paper, we propose the first multi-model prediction (MoP) method for image set compression to significantly reduce inter image redundancy. Unlike the previous prediction methods, our MoP enhances the correlation between images using feature-based geometric multi-model fitting. Based on estimated geometric models, multiple deformed prediction images are generated to reduce geometric distortions in different image regions. The block-based adaptive motion compensation is then adopted to further eliminate local variances. Experimental results demonstrate the advantage of our approach, especially for images with complicated scenes and geometric relationships.

Traditional communication systems usually suffer from the threshold effect when channel signal-to-noise ratio (CSNR) fluctuates unpredictably in wireless and mobile scenarios. The SoftCast scheme, however, provides graceful quality transition in wide CSNR range. In SoftCast, input image is decorrelated by a transform and modulated directly to a dense constellation for transmission, leaving out the conventional quantization, entropy coding and channel coding. A key point of SoftCast is that the transmission power needs to be allocated among the transform coefficients unequally, according to the energy of coefficients. Importantly, the energy diversity used to guide power allocation should be shared between the sender and the receiver for correct decoding. This paper addresses the power distortion optimization problem, introducing a new adaptive chunk division scheme to describe the energy diversity among coefficients. A concrete algorithm is developed to determine the chunk boundaries that achieve optimal transmission power usage. Experimental results show that the proposed scheme can improve the performance of the original SoftCast by 4~8dB using a smaller number of chunks.

This paper proposes a novel framework called DCast for distributed video coding and transmission over wireless networks, which is different from existing distributed schemes in three aspects. First, coset quantized DCT coefficients and motion data are directly delivered to the channel coding layer without syndrome or entropy coding. Second, transmission power is directly allocated to coset data and motion data according to their distributions and magnitudes without forward error correction. Third, these data are transformed by Hadamard and then directly mapped using a dense constellation (64K-QAM) for transmission without Gray coding. One of the most important properties in this framework is that the coding and transmission rate is fixed and distortion is minimized by allocating the transmission power. Thus, we further propose a power distortion optimization algorithm to replace the traditional rate distortion optimization. This framework avoids the annoying cliff effect caused by the mismatch between transmission rate and channel condition. In multicast, each user can get approximately the best quality matching its channel condition. Our experiment results show that the proposed DCast outperforms the typical solution using H.264 over 802.11 up to 8 dB in video PSNR in video broadcast. Even in video unicast, the proposed DCast is still comparable to the typical solution.