Human motion synthesis in 3D scenes relies heavily on scene comprehension, while current methods focus mainly on scene structure but ignore the semantic understanding. In this paper, we propose a human motion synthesis framework that take an unified Scene Semantic Occupancy (SSO) for scene representation, termed SSOMotion. We design a bi-directional tri-plane decomposition to derive a compact version of the SSO, and scene semantics are mapped to an unified feature space via CLIP encoding and shared linear dimensionality reduction. Such strategy can derive the fine-grained scene semantic structures while significantly reduce redundant computations. We further take these scene hints and movement direction derived from instructions for motion control via frame-wise scene query. Extensive experiments and ablation studies conducted on cluttered scenes using ShapeNet furniture, as well as scanned scenes from PROX and Replica datasets, demonstrate its cutting-edge performance while validating its effectiveness and generalization ability.

Scene-aware motion synthesis has been widely researched recently due to its numerous applications. Prevailing methods rely heavily on paired motion-scene data, while it is difficult to generalize to diverse scenes when trained only on a few specific ones. Thus, we propose a unified framework, termed Diffusion Implicit Policy (DIP), for scene-aware motion synthesis, where paired motion-scene data are no longer necessary. In this paper, we disentangle human-scene interaction from motion synthesis during training, and then introduce an interaction-based implicit policy into motion diffusion during inference. Synthesized motion can be derived through iterative diffusion denoising and implicit policy optimization, thus motion naturalness and interaction plausibility can be maintained simultaneously. For long-term motion synthesis, we introduce motion blending in joint rotation power space. The proposed method is evaluated on synthesized scenes with ShapeNet furniture, and real scenes from PROX and Replica. Results show that our framework presents better motion naturalness and interaction plausibility than cutting-edge methods. This also indicates the feasibility of utilizing the DIP for motion synthesis in more general tasks and versatile scenes.

Reconstructing dynamic scenes has long been a challenging task in 3D vision. Previous mainstream methods based on 3D Gaussian Splatting typically employ a single deformation field to directly model spatiotemporal changes. However, such one-step deformation struggles to capture diverse and complex motion patterns. To address this limitation, we propose decomposing the one-step deformation into a multi-step process, where each step is represented by a deformation layer. Additionally, we introduce a weight prediction mechanism for each layer to control the extent of deformation at every step. We provide two types of deformation layers based on implicit and explicit approaches. Moreover, while the deformation layer is time-conditioned, the Gaussians' behavior may still be influenced by their time-invariant properties. Therefore, we propose a fully time-agnostic scale modulation block to modulate the scaling changes of Gaussians. Extensive experiments on D-NeRF, Neu3D, and HyperNeRF demonstrate that our method achieves state-of-the-art performance.

Open-vocabulary semantic segmentation (OVSS) employs pixel-level vision-language alignment to associate category-related prompts with corresponding pixels. A key challenge is enhancing the multimodal dense prediction capability, specifically this pixel-level multimodal alignment. Although existing methods achieve promising results by leveraging CLIP’s vision-language alignment, they rarely investigate the performance boundaries of CLIP for dense prediction from an interpretability mechanisms perspective. In this work, we systematically investigate CLIP's internal mechanisms and identify a critical phenomenon: analogous to human distraction, CLIP diverts significant attention resources from target regions to irrelevant tokens. Our analysis reveals that these tokens arise from dimension-specific over-activation; filtering them enhances CLIP's dense prediction performance. Consequently, we propose Refocusing CLIP (RF-CLIP), a training-free approach that emulates human distraction-refocusing behavior to redirect attention from distraction tokens back to target regions, thereby refining CLIP's multimodal alignment granularity. Our method achieves SOTA performance on eight benchmarks while maintaining high inference efficiency.

3D Visual Grounding (3DVG) aims to localize the referent of natural language referring expressions through two core tasks: Referring Expression Comprehension (3DREC) and Segmentation (3DRES). While existing methods achieve high accuracy in simple, single-object scenes, they suffer from severe performance degradation in complex, multi-object scenes that are common in real-world settings, hindering practical deployment. Existing methods face two key challenges in complex, multi-object scenes: inadequate parsing of implicit localization cues critical for disambiguating visually similar objects, and ineffective suppression of dynamic spatial interference from co-occurring objects, resulting in degraded grounding accuracy. To address these challenges, we propose PC-CrossDiff, a unified dual-task framework with a dual-level cross-modal differential attention architecture for 3DREC and 3DRES. Specifically, the framework introduces: (i) Point-Level Differential Attention (PLDA) modules that apply bidirectional differential attention between text and point clouds, adaptively extracting implicit localization cues via learnable weights to improve discriminative representation; (ii) Cluster-Level Differential Attention (CLDA) modules that establish a hierarchical attention mechanism to adaptively enhance localization-relevant spatial relationships while suppressing ambiguous or irrelevant spatial relations through a localization-aware differential attention block. To address the scale disparity and conflicting gradients in joint 3DREC–3DRES training, we propose L_DGTL, a unified loss function that explicitly reduces multi-task crosstalk and enables effective parameter sharing across tasks. Our method achieves state-of-the-art performance on the ScanRefer, NR3D, and SR3D benchmarks. Notably, on the Implicit subsets of ScanRefer, it improves the [email protected] score by +10.16% for the 3DREC task, highlighting its strong ability to parse implicit spatial cues.

Existing end-to-end approaches of robotic manipulation often lack generalization to unseen objects or tasks due to limited data and poor interpretability. While recent Multimodal Large Language Models (MLLMs) demonstrate strong commonsense reasoning, they struggle with geometric and spatial understanding required for pose prediction. In this paper, we propose RobMRAG, a 3D Gaussian Splatting-Enhanced Multimodal Retrieval-Augmented Generation (MRAG) framework for zero-shot robotic manipulation. Specifically, We construct a multi-source manipulation knowledge base containing object contact frames, task completion frames, and pose parameters. During inference, a Hierarchical Multimodal Retrieval module first employs hybrid semantic search to find task-relevant object prototypes, then selects the geometrically closest reference example based on pixel-level similarity and Instance Matching Distance (IMD). We further introduce a 3D-Aware Pose Refinement module based on 3D Gaussian Splatting into the MRAG framework, which aligns the pose of the reference object to the target object in 3D space. The aligned results are reprojected onto the image plane and used as input to the MLLM to enhance the generation of the final pose parameters. Extensive experiments show that on a test set containing 30 categories of household objects, our method improves the success rate by 7.76% compared to the best-performing zero-shot baseline under the same setting, and by 6.54% compared to the state-of-the-art supervised baseline. Our results validate that RobMRAG effectively bridges the gap between high-level semantic reasoning and low-level geometric execution, enabling robotic systems that generalize to unseen objects while remaining inherently interpretable.

Zhihao Luo, Wentao Yan, Jingyu Gong, Min Wang, Zhizhong Zhang, Xuhong Wang, Yuan Xie, Xin Tan. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Real-world industrial scenarios have become increasingly dynamic, with new product types, defect patterns, and operational modes emerging rapidly. In such a context, the one-for-more paradigm enables the use of a single model to economically and continually adapt to evolving distributions or patterns, positioning it as a key component in modern Industrial AI systems. This article proposes a novel one-for-more anomaly detection framework designed to identify anomalies across expanding product lines. The framework incorporates two model-agnostic techniques: instance-aware prompt tuning (IPT) and gradient-aware parameter decoupling (GPD). Our approach is built upon a reconstruction-based vision transformer (ViT) encoder–decoder architecture. IPT addresses the domain gap between pretrained models and industrial data by leveraging an instance-level prompt and a shared memory mechanism, which helps the pretrained model retain previously learned patterns. GPD selectively updates network parameters based on the gradient’s impact on prior tasks, employing orthogonal gradient projection to further minimize interference. In addition, we introduce a new dataset to simulate the one-for-more industrial scenario. Extensive experiments on MVTec and our proposed dataset demonstrate that our framework achieves the state-of-the-art performance across various continual learning settings, significantly outperforming existing methods, particularly in multistep incremental scenarios.

Industrial few-shot anomaly detection (FSAD) requires identifying various abnormal states by leveraging as few normal samples as possible (abnormal samples are unavailable during training). However, current methods often require training a separate model for each category, leading to increased computation and storage overhead. Thus, designing a unified anomaly detection model that supports multiple categories remains a challenging task, as such a model must recognize anomalous patterns across diverse objects and domains. To tackle these challenges, this paper introduces FocusPatch AD, a unified anomaly detection framework based on vision-language models, achieving anomaly detection under few-shot multi-class settings. FocusPatch AD links anomaly state keywords to highly relevant discrete local regions within the image, guiding the model to focus on cross-category anomalies while filtering out background interference. This approach mitigates the false detection issues caused by global semantic alignment in vision-language models. We evaluate the proposed method on the MVTec, VisA, and Real-IAD datasets, comparing them against several prevailing anomaly detection methods. In both image-level and pixel-level anomaly detection tasks, FocusPatch AD achieves significant gains in classification and localization performance, demonstrating excellent generalization and adaptability.

Unsupervised visible infrared person re-identification (USVI-ReID) is a challenging retrieval task that retrieves cross-modality pedestrian images without using any label information. In this task, the large cross-modality variance makes it difficult to generate reliable cross-modality labels, and the lack of annotations also provides additional difficulties for learning modality-invariant features. To facilitate this unsupervised cross-modal learning, we begin by leveraging the information contained in the cross-modality input and its predicted label. Aiming to minimize information loss, we optimize the model by incorporating entropy minimization, uniform label distribution, and cross-modality matching. In our approach, we design a loop iterative training strategy alternating between model training and cross-modality matching, where a uniform prior guided optimal transport assignment is proposed to select matched visible and infrared prototypes. This matching information is then utilized to minimize the intra- and cross-modality entropy. As a result, our model can gradually self-learn useful information, enabling it to generate discriminative representations for unlabeled cross-modal data. Extensive experimental results on benchmarks demonstrate the effectiveness of our method, e.g., 69.4% and 89.4% of Rank-1 accuracy on SYSU-MM01 and RegDB without any annotations. The code will be released soon.

Point cloud semantic segmentation, which involves assigning a category for each point, is a crucial task in autonomous driving and intelligent transportation systems. Due to the inherently unordered and irregular nature of point clouds, learning robust features that accurately capture real-world distributions from point coordinates and other attributes remains challenging. Following the pioneering work of PointNet, current 3D deep neural networks process point coordinates alongside other attributes without fully exploiting the implicit class prior information embedded in spatial information. In this work, we first conduct a pilot study to evaluate how current 3D networks utilize point coordinates and validate the presence of implicit class priors within them. Subsequently, we design a robust Position-to-Physics (P2P) fusion strategy that learns adaptive weights to dynamically incorporate implicit class priors present in point coordinates into point features. Moreover, we design a dual-branch network architecture and propose a triplet loss to further enhance the adaptive fusion process. Extensive experiments demonstrate that decoupling position attributes from physics attributes facilitates the extraction and utilization of implicit class priors. Our proposed modules consistently improve segmentation performance across various networks and datasets, demonstrating their generalizability and effectiveness.

The semantically interactive radiance field has always been an appealing task for its potential to facilitate user-friendly and automated real-world 3D scene understanding applications. However, it is a challenging task to achieve high quality, efficiency and zero-shot ability at the same time with semantics in radiance fields. In this work, we present FastLGS, an approach that supports real-time open-vocabulary query within 3D Gaussian Splatting (3DGS) under high resolution. We propose the semantic feature grid to save multi-view CLIP features which are extracted based on Segment Anything Model (SAM) masks, and map the grids to low dimensional features for semantic field training through 3DGS. Once trained, we can restore pixel-aligned CLIP embeddings through feature grids from rendered features for open-vocabulary queries. Comparisons with other state-of-the-art methods prove that FastLGS can achieve the first place performance concerning both speed and accuracy, where FastLGS is 98 times faster than LERF, 4 times faster than LangSplat and 2.5 times faster than LEGaussians. Meanwhile, experiments show that FastLGS is adaptive and compatible with many downstream tasks, such as 3D segmentation and 3D object inpainting, which can be easily applied to other 3D manipulation systems.

Domain Generalized Semantic Segmentation (DGSS) aims to utilize segmentation model training on known source domains to make predictions on unknown target domains. Currently, there are two network architectures: one based on Convolutional Neural Networks (CNNs) and the other based on Visual Transformers (ViTs). However, both CNN-based and ViT-based DGSS methods face challenges: the former lacks a global receptive field, while the latter requires more computational demands. Drawing inspiration from State Space Models (SSMs), which not only possess a global receptive field but also maintain linear complexity, we propose SSM-based method for achieving DGSS. In this work, we first elucidate why does mask make sense in SSM-based DGSS and propose our mask learning mechanism. Leveraging this mechanism, we present our Mask Vision Mamba network (MaskViM), a model for SSM-based DGSS, and design our mask loss to optimize MaskViM. Our method achieves superior performance on four diverse DGSS setting, which demonstrates the effectiveness of our method.

We propose DrivingForward, a feed-forward Gaussian Splatting model that reconstructs driving scenes from flexible surround-view input. Driving scene images from vehicle-mounted cameras are typically sparse, with limited overlap, and the movement of the vehicle further complicates the acquisition of camera extrinsics. To tackle these challenges and achieve real-time reconstruction, we jointly train a pose network, a depth network, and a Gaussian network to predict the Gaussian primitives that represent the driving scenes. The pose network and depth network determine the position of the Gaussian primitives in a self-supervised manner, without using depth ground truth and camera extrinsics during training. The Gaussian network independently predicts primitive parameters from each input image, including covariance, opacity, and spherical harmonics coefficients. At the inference stage, our model can achieve feed-forward reconstruction from flexible multi-frame surround-view input. Experiments on the nuScenes dataset show that our model outperforms existing state-of-the-art feed-forward and scene-optimized reconstruction methods in terms of reconstruction.

Source-Free Domain Adaptation (SFDA) aims to transfer a pre-trained source model to the unlabeled target domain without accessing the source data, thereby effectively solving labeled data dependency and domain shift problems. However, the SFDA setting faces a bottleneck due to the absence of supervisory information. To mitigate this problem, Active Learning (AL) is introduced to combine with SFDA, endeavoring to actively label a small set of the most high-quality target points so that models with satisfactory performance can be obtained at an acceptable cost. Nevertheless, several issues remain unresolved, namely when to query new labels during training, what kind of samples deserve labeling to ensure rich information, and where the labels should be distributed to guarantee diversity. Thus we elaborate OmniQuery to omnibearing address the “When, What, and Where” problems about active points querying in source-free domain adaptation for cross-modal 3D semantic segmentation. The method consists of three main components: Query Decider, Point Ranker, and Budget Slicer. The Query Decider determines the optimal timing to query new points by fitting the validation curves during training. The Point Ranker nominates points for annotation by calculating the ambiguity of neighboring points in the feature space. The Budget Slicer allocates the annotation quota, i.e., labeling percentage of the point cloud, to different semantic regions by utilizing the advanced 2D semantic segmentation capabilities of the Segment Anything Model (SAM). Extensive experiments demonstrate the effectiveness of our proposed method, achieving up to 99.64% of fully supervised performance with only 3% of labels, and consistently outperforming comparison methods across various scenarios.

With the rise of generative models, there is a growing interest in unifying all tasks within a generative framework. Anomaly detection methods also fall into this scope and utilize diffusion models to generate or reconstruct normal samples when given arbitrary anomaly images. However, our study found that the diffusion model suffers from severe "faithfulness hallucination" and "catastrophic forgetting", which can’t meet the unpredictable pattern increments. To mitigate the above problems, we propose a continual diffusion model that uses gradient projection to achieve stable continual learning. Gradient projection deploys a regularization on the model updating by modifying the gradient towards the direction protecting the learned knowledge. But as a double-edged sword, it also requires huge memory costs brought by the Markov process. Hence, we propose an iterative singular value decomposition method based on the transitive property of linear representation, which consumes tiny memory and incurs almost no performance loss. Finally, considering the risk of "over-fitting" to normal images of the diffusion model, we propose an anomaly-masked network to enhance the condition mechanism of the diffusion model. For continual anomaly detection, ours achieves first place in 17/18 settings on MVTec and VisA. Code is available at https://github.com/FuNz-0/One-for-More

The nearest prototypical classifier faces challenges of semantic drift and prototype interference. Previous methods address these issues using data rehearsal and contrastive learning, but these approaches incur high memory costs and slow convergence. In this paper, we propose a novel prototypical minimum distance loss, along with a two-stage training pipeline, to mitigate prototype interference with low memory overhead and fast convergence. Leveraging task-specific prompts and a key-query mechanism, we significantly reduce semantic drift. Additionally, we introduce a continual exponential moving average to enhance model stability and minimize forgetting. Notably, our method is rehearsal-free and avoids generation processes, simplifying training and further reducing memory usage. We validate our approach on four challenging class-incremental learning datasets, achieving significant improvements over state-of-the-art methods.

Recent advancements in pre-trained models have enhanced performance on downstream tasks due to their strong generalizability. Despite this, models fine-tuned continually often face challenges such as catastrophic forgetting and loss of generalization. To address these issues, we propose a novel approach that utilizes distinct Low-Rank Adaptation (LoRA) modules for each task. These modules parameter-efficient, and integrated across tasks to ensure the model maintains strong performance on both old and new classes. Additionally, we investigate semantic relationships between class prototypes to effectively reconstruct old prototypes in the context of new tasks. Our experiments demonstrate that this method significantly outperforms baseline approaches across various class-incremental learning benchmarks, offering an efficient and effective solution for mitigating forgetting and preserving model performance.

Cross-modal fusion, which leverages images to enhance 3D semantic segmentation, has demonstrated significant effectiveness due to the complementary nature of heterogeneous data. However, existing approaches are limited to pinhole images, leaving fisheye images largely unexplored. In this paper, we introduce the LiDAR-Fisheye Fusion Network (LFNet), a dual-transformer architecture designed for cross-modal fusion (CMF) across hierarchical multi-scale layers. The 3D Transformer extracts point-level features from LiDAR data, while the pre-trained 2D Transformer extracts patch-level features from fisheye images.The CMF module comprises two key components: Local Fusion (LoF) and Global Fusion (GoF). The LoF module interpolates patch-level features to pixel-level for accurate feature alignment and computes precise point-to-pixel mappings for gated fusion. Meanwhile, the GoF module enables points to capture a holistic understanding of the scene via a cross-modal attention mechanism. Experimental results highlight the potential of fisheye images as a promising modality to complement LiDAR data in 3D semantic segmentation. The code will be available at https://github.com/wjzhang642/LFNet.

Continual Instruction Tuning (CIT) is adopted to continually instruct Large Models to follow human intent data by data. It is observed that existing gradient update would heavily destroy the performance on previous datasets during CIT process. Instead, Exponential Moving Average (EMA), owns the ability to trace previous parameters, which can aid in decreasing forgetting. Nonetheless, its stable balance weight fails to deal with the ever-changing datasets, leading to the out-of-balance between plasticity and stability. In this paper, we propose a general continual instruction tuning framework to address the challenge. Starting from the trade-off prerequisite and EMA update, we propose the plasticity and stability ideal condition. Based on Taylor expansion in the loss function, we find the optimal balance weight can be automatically determined by the gradients and learned parameters. Therefore, we propose a stable-plasticity balanced coefficient to avoid knowledge interference. Based on the semantic similarity of the instructions, we can determine whether to retrain or expand the training parameters and allocate the most suitable parameters for the testing instances. Extensive experiments across multiple continual instruction tuning benchmarks demonstrate that our approach not only enhances anti-forgetting capabilities but also significantly improves overall continual tuning performance. Our code is available at https://github.com/JingyangQiao/CoIN.

In this paper, we present the first pinhole-fisheye framework for heterogeneous multi-view depth estimation, PFDepth. Our key insight is to exploit the complementary characteristics of pinhole and fisheye imagery (undistorted vs. distorted, small vs. large FOV, far vs. near field) for joint optimization. PFDepth employs a unified architecture capable of processing arbitrary combinations of pinhole and fisheye cameras with varied intrinsics and extrinsics. Within PFDepth, we first explicitly lift 2D features from each heterogeneous view into a canonical 3D volumetric space. Then, a core module termed Heterogeneous Spatial Fusion is designed to process and fuse distortion-aware volumetric features across overlapping and non-overlapping regions. Additionally, we subtly reformulate the conventional voxel fusion into a novel 3D Gaussian representation, in which learnable latent Gaussian spheres dynamically adapt to local image textures for finer 3D aggregation. Finally, fused volume features are rendered into multi-view depth maps. Through extensive experiments, we demonstrate that PFDepth sets a state-of-the-art performance on KITTI-360 and RealHet datasets over current mainstream depth networks. To the best of our knowledge, this is the first systematic study of heterogeneous pinhole-fisheye depth estimation, offering both technical novelty and valuable empirical insights.

Open-vocabulary semantic segmentation (OVSS) conducts pixel-level classification via text-driven alignment, where the domain discrepancy between base category training and open-vocabulary inference poses challenges in discriminative modeling of latent unseen category. To address this challenge, existing vision-language model (VLM)-based approaches demonstrate commendable performance through pre-trained multi-modal representations. However, the fundamental mechanisms of latent semantic comprehension remain underexplored, making the bottleneck for OVSS. In this work, we initiate a probing experiment to explore distribution patterns and dynamics of latent semantics in VLMs under inductive learning paradigms. Building on these insights, we propose X-Agent, an innovative OVSS framework employing latent semantic-aware ''agent'' to orchestrate cross-modal attention mechanisms, simultaneously optimizing latent semantic dynamic and amplifying its perceptibility. Extensive benchmark evaluations demonstrate that X-Agent achieves state-of-the-art performance while effectively enhancing the latent semantic saliency.

3D Visual Grounding (3DVG) aims to localize objects in 3D scenes using natural language descriptions. Although supervised methods achieve higher accuracy in constrained settings, zero-shot 3DVG holds greater promise for real-world applications since eliminating scene-specific training requirements. However, existing zero-shot methods face challenges of spatial-limited reasoning due to reliance on single-view localization, and contextual omissions or detail degradation. To address these issues, we propose SeqVLM, a novel zero-shot 3DVG framework that leverages multi-view real-world scene images with spatial information for target object reasoning. Specifically, SeqVLM first generates 3D instance proposals via a 3D semantic segmentation network and refines them through semantic filtering, retaining only semantic-relevant candidates. A proposal-guided multi-view projection strategy then projects these candidate proposals onto real scene image sequences, preserving spatial relationships and contextual details in the conversion process of 3D point cloud to images. Furthermore, to mitigate VLM computational overload, we implement a dynamic scheduling mechanism that iteratively processes sequances-query prompts, leveraging VLM's cross-modal reasoning capabilities to identify textually specified objects. Experiments on the ScanRefer and Nr3D benchmarks demonstrate state-of-the-art performance, achieving [email protected] scores of 55.6% and 53.2%, surpassing previous zero-shot methods by 4.0% and 5.2%, respectively, which advance 3DVG toward greater generalization and real-world applicability.

Multi-modal test-time adaptation (TTA) for 3D semantic segmentation has increasingly become a research hotspot due to its ability to address label dependency and enable rapid adaptation. Existing methods rely on learnable extra components to mitigate reliability bias, however, learning-based approaches in TTA scenarios often lack sufficient training. Moreover, most existing approaches update only normalization layers in the teacher-student framework, which limits their ability to model domain shifts. To overcome these limitations, we propose PLATO-TTA, a novel multi-modal TTA method for 3D semantic segmentation leveraging the native stability in robust prototypes and adaptive tuning of critical teacher-student parameters. The approach contains three key components: Prototype-Guided Pseudo-Labeling (PGPL), Consistency Based Backtracking (CBB), and Domain Specific Updating (DSU). PGPL reduces reliability bias by constructing pseudo-source domain prototypes and computing modality fusion weights based on domain discrepancies. CBB updates all student model parameters while preventing catastrophic forgetting through a parameter backtracking mechanism. DSU selectively updates the teacher model using only domain-specific parameters from the student model, ensuring rapid adaptation and stable guidance. Extensive experiments demonstrate the effectiveness of PLATO-TTA, bringing a 6.3% gain to the SynthiatoSemanticKITTI scenario with severe reliability bias and significant domain discrepancy, and achieve state-of-the-art performance across various domain adaptation scenarios.

Current outdoor point-cloud segmentation methods typically formulate semantic segmentation as a per-point/voxel-classification task. Although this strategy is straightforward because it classifies each point directly, it ignores the overall relationship of the category. As an alternative paradigm, mask classification decouples category classification from region localization, allowing the model to better capture overall category relationships. In this paper, we propose a novel approach called the point mask transformer (PMFormer), which transforms the semantic segmentation of point clouds from per-point classification to mask classification using a transformer architecture. The proposed model comprises a 3D backbone, transformer decoder, and segmentation head that predicts a series of binary masks, each associated with a global class label. Furthermore, to accommodate the unique characteristics of large and sparse outdoor point-cloud scenes, we propose three enhancements for the integration of point-cloud data with the transformer: MaskMix, 3D position encoding, and attention weights. We evaluate our model using the SemanticKITTI and nuScenes datasets. Our experimental results show that the proposed method outperforms state-of-the-art semantic segmentation approaches.

Parameter-efficient tunings (PETs) have demonstrated impressive performance and promising perspectives in training large models, while they are still confronted with a common problem: the trade-off between learning new content and protecting old knowledge, leading to zero-shot generalization collapse, and cross-modal hallucination. In this paper, we reformulate Adapter, LoRA, Prefix-tuning, and Prompt-tuning from the perspective of gradient projection, and first propose a unified framework called Parameter Efficient Gradient Projection (PEGP). We introduce orthogonal gradient projection into different PET paradigms and theoretically demonstrate that the orthogonal condition for the gradient can effectively resist forgetting even for large-scale models. It therefore modifies the gradient towards the direction that has less impact on the old feature space, with less extra memory space and training time. We extensively evaluate our method with different backbones, including ViT and CLIP, on diverse datasets, and experiments comprehensively demonstrate its efficiency in reducing forgetting in class, online class, domain, task, and multi-modality continual settings.

Facial micro-expression recognition (MER) is a challenging problem, due to transient and subtle micro-expression (ME) actions. Most existing methods depend on hand-crafted features, key frames like onset, apex, and offset frames, or deep networks limited by small-scale and low-diversity datasets. In this paper, we propose an end-to-end micro-action-aware deep learning framework with advantages from transformer, graph convolution, and vanilla convolution. In particular, we propose a novel F5C block composed of fully-connected convolution and channel correspondence convolution to directly extract local-global features from a sequence of raw frames, without the prior knowledge of key frames. The transformer-style fully-connected convolution is proposed to extract local features while maintaining global receptive fields, and the graph-style channel correspondence convolution is introduced to model the correlations among feature patterns. Moreover, MER, optical flow estimation, and facial landmark detection are jointly trained by sharing the local-global features. The two latter tasks contribute to capturing facial subtle action information for MER, which can alleviate the impact of insufficient training data. Extensive experiments demonstrate that our framework (i) outperforms the state-of-the-art MER methods on CASME II, SAMM, and SMIC benchmarks, (ii) works well for optical flow estimation and facial landmark detection, and (iii) can capture facial subtle muscle actions in local regions associated with MEs.

Unlike Conventional Zero-Shot Learning (CZSL) which only focuses on the recognition of unseen classes by using a classifier trained on seen classes and semantic embeddings, Generalized Zero-Shot Learning (GZSL) requires a classifier trained on seen classes to recognize objects from both seen and unseen classes. To tackle this problem, feature generative-based models have been proposed to synthesize visual features for unseen classes conditioned on their semantic descriptors. However, they treat these semantic descriptors as independent individuals without exploring their structural relations among categories. We propose a novel approach, dubbed Relation Extrapolation based feature generation for GZSL (RE-GZSL), which generates features of unseen classes by borrowing some features that are extrapolated from seen classes based on semantic relations. In RE-GZSL, a visual-semantic relations alignment loss and an instance-prototype contrastive loss are presented to align visual relations with semantic relations. To maintain the information of the visual features before and after the alignment, a discrimination preservation loss is further introduced. Besides, a feature mixing module is built to synthesize features for unseen classes, which are more realistic and tightly related to seen classes. Experimental results demonstrate that RE-GZSL outperforms competitors on four benchmark datasets. Comprehensive ablation studies and analyses are provided to dissect what factors led to this success. Code is available at:https://github.com/Barcaaaa/RE-GZSL.

In cross-modal unsupervised domain adaptation, a model trained on source-domain data (e.g., synthetic) is adapted to target-domain data (e.g., real-world) without access to target annotation. Previous methods seek to mutually mimic cross-modal outputs in each domain, which enforces a class probability distribution that is agreeable in different domains. However, they overlook the complementarity brought by the heterogeneous fusion in cross-modal learning. In light of this, we propose a novel fusion-then-distillation (FtD++) method to explore cross-modal positive distillation of the source and target domains for 3D semantic segmentation. FtD++ realizes distribution consistency between outputs not only for 2D images and 3D point clouds but also for source-domain and augment-domain. Specially, our method contains three key ingredients. First, we present a model-agnostic feature fusion module to generate the cross-modal fusion representation for establishing a latent space. In this space, two modalities are enforced maximum correlation and complementarity. Second, the proposed cross-modal positive distillation preserves the complete information of multi-modal input and combines the semantic content of the source domain with the style of the target domain, thereby achieving domain-modality alignment. Finally, cross-modal debiased pseudo-labeling is devised to model the uncertainty of pseudo-labels via a self-training manner. Extensive experiments report state-of-the-art results on several domain adaptive scenarios under unsupervised and semi-supervised settings. Code is available athttps://github.com/Barcaaaa/FtD-PlusPlus

Unsupervised Visible-Infrared Person Re-Identification (USVI-ReID) aims to match visible and infrared person images without relying on prior annotations. Recently, unsupervised contrastive learning methods have become the mainstream approach for USVI-ReID, leveraging clustering algorithms to generate pseudo-labels. However, these methods often suffer from inherent noisy pseudo-labels, which significantly hinders their performance. To address this challenge, we propose a Adaptive Pseudo-label Purification and Debiasing (APPD) framework for USVI-ReID, which is designed to calibrate noisy pseudo-labels and dynamically detects clean pseudo-labels, thereby enhancing the model’s performance and reliability. Specifically, we propose an Adaptive Pseudo-label Calibration and Division (APCD) module, which calibrates noisy pseudo-labels by assessing their reliability and divides pseudo-labels into clean and noisy subsets, ensuring a more focused and accurate learning process. Based on the calibrated pseudo-labels, we develop an Optimal Transport Prototype Matching (OTPM) module to establish robust cross-modality correspondences. For clean pseudo-labels, we propose a Debiased Memory Hybrid Learning (DMHL) module, which jointly captures modality-specific and modality-invariant information while addressing sampling bias to enhance feature representation. To effectively utilize noisy pseudo-labels, we introduce a Neighbor Relation Learning (NRL) module that mitigates intra-class variations by exploring neighbor relationships in the feature space. Comprehensive experiments conducted on two widely recognized USVI-ReID benchmarks demonstrate that APPD achieves state-of-the-art performance, significantly outperforming existing methods. The source code will be made available at https://github.com/XiangboYin/RPNR.

3D occupancy perception holds a pivotal role in recent vision-centric autonomous driving systems by converting surround-view images into integrated geometric and semantic representations within dense 3D grids. Nevertheless, current models still encounter two main challenges: modeling depth accurately in the 2D-3D view transformation stage, and overcoming the lack of generalizability issues due to sparse LiDAR supervision. To address these issues, this paper presents GEOcc, a Geometric-Enhanced Occupancy network tailored for vision-only surround-view perception. Our approach is three-fold: 1) Integration of explicit lift-based depth prediction and implicit projection-based transformers for depth modeling, enhancing the density and robustness of view transformation. 2) Utilization of mask-based encoder-decoder architecture for fine-grained semantic predictions; 3) Adoption of context-aware self-training loss functions in the pertaining stage to complement LiDAR supervision, involving the re-rendering of 2D depth maps from 3D occupancy features and leveraging image reconstruction loss to obtain denser depth supervision besides sparse LiDAR ground-truths. Our approach achieves State-of-the-Art performance on the Occ3D-nuScenes dataset with the least image resolution needed and the most weightless image backbone compared with current models, marking an improvement of 3.3% due to our proposed contributions. Comprehensive experimentation also demonstrates the consistent superiority of our method over baselines and alternative approaches. Our code is available athttps://github.com/world-executed/GEOcc.git

Alignment between the food images and the corresponding recipes is an emerging cross-modal representation learning task. In this task, the recipes are composed of three components, i.e., food title, ingredient lists, and cooking instructions, which require a fine-grained alignment between the features of the two modalities. Existing methods usually aggregate the recipes into global embeddings and then align them with the global image embeddings. Meanwhile, semantic classification is frequently used in these methods to regularize the embeddings of the two modalities. While these methods are efficient, there remain two problems: (1) Forcing the alignment between the global images and recipes embeddings may result in losing the component-specific information. (2) The high diversity of food appearance leads to high uncertainty in the semantic classification of food images and recipes. To solve these problems, we propose a Fine-grained Prompting and Alignment (FPA) model to enhance the feature extraction and bring more component-specific information for fine-grained alignment. Furthermore, to regularize the semantic information contained in the cross-modal features, we design an Evidential Semantic Consistency (ESC) loss to keep the cross-modal semantic consistency. We have conducted comprehensive experiments on the benchmark dataset Recipe1M and the state-of-the-art results on the cross-modal recipe retrieval task demonstrate the effectiveness of our method.

Weakly supervised point cloud semantic segmentation is an increasingly active topic, because fully supervised learning acquires well-labeled point clouds and entails high costs. The existing weakly supervised methods either need meticulously designed data augmentation for self-supervised learning or ignore the negative effects of learning on pseudolabel noises. In this article, by designing different granularity of cross-cloud structures, we propose a cross-cloud consistency method for weakly supervised point cloud semantic segmentation which forms the expectation-maximum (EM) framework. Benefiting from the cross-cloud constraints, our method allows effective learning alternatively between refining pseudolabels and updating network parameters. Specifically, in E-step, we propose a pseudolabel selecting (PLS) strategy based on cross subcloud consistency, improving the credibility of selected pseudolabels explicitly. In M-step, a cross-scene contrastive regularization enforces cross-scene prototypes with the same label in different scenes to be more similar, while keeping prototypes with different labels to be a clear margin, reducing the noise fitting. Finally, we give some insight into the optimization of our method in the EM theoretical way. The proposed method is evaluated on three challenging datasets, where experimental results demonstrate that our method significantly outperforms state-of-the-art weakly supervised competitors. Our code is available online: https://github.com/Yachao-Zhang/Cross-Cloud-Consistency.

Humans can quickly learn new concepts with limited experience, while not forgetting learned knowledge. Such ability in machine learning is referred to as few-shot class-incremental learning (FSCIL). Although some methods try to solve this problem by putting similar efforts to prevent forgetting and promote learning, we find existing techniques do not give enough importance to the new category as new training samples are rather rare. In this article, we propose a new biased-to-unbiased rectification method, which introduces a trainable transition matrix to mitigate the prediction discrepancy between the old classes and the new classes. This transition matrix is to be diagonally dominated, normalized, and differentiable with new-knowledge-preferred prior, to solving the strong bias between heavy old knowledge and limited new knowledge. Hence, we can achieve a balanced solution between learning new concepts and preventing catastrophic forgetting by giving new classes more chances. Extensive experiments on miniImagenet, CIFAR100, and CUB200 demonstrate that our method outperforms the latest state-of-the-art methods by 1.1%, 1.44%, and 2.08%, respectively.

Due to the unsatisfactory performance of supervised methods on unpaired real-world scans, point cloud completion via cross-domain adaptation has recently drawn growing attention. Nevertheless, previous approaches only focus on alleviating the distribution shift through domain alignment, resulting in massive information loss of real-world domain data. To tackle this issue, we propose a dual mixup-induced consistency regularization to integrate both source and target domain to improve robustness and generalization capability. Specifically, we mix up virtual and real-world shapes in the input and latent feature space respectively, and then regularize the completion network by forcing two kinds of mixed completion predictions to be consistent. To further adapt to each instance within the real-world domain, we design a novel density-aware refiner to utilize local context information to preserve the fine-grained details and remove noise or outliers for coarse completion. Extensive experiments on real-world scans and our synthetic unpaired datasets demonstrate the superiority of our method over existing state-of-the-art approaches.

As the exorbitant expense of labeling autopilot datasets and the growing trend of utilizing unlabeled data, semi-supervised segmentation on point clouds becomes increasingly imperative. Intuitively, finding out more ``unspoken words'' (i.e., latent instance information) beyond the label itself should be helpful to improve performance. In this paper, we discover two types of latent labels behind the displayed label embedded in LiDAR and image data. First, in the LiDAR Branch, we propose a novel augmentation, Cylinder-Mix, which is able to augment more yet reliable samples for training. Second, in the Image Branch, we propose the Instance Position-scale Learning (IPSL) Module to learn and fuse the information of instance position and scale, which is from a 2D pre-trained detector and a type of latent label obtained from 3D to 2D projection. Finally, the two latent labels are embedded into the multi-modal panoptic segmentation network. The ablation of the IPSL module demonstrates its robust adaptability, and the experiments evaluated on SemanticKITTI and nuScenes demonstrate that our model outperforms the state-of-the-art method, LaserMix.

It is well-known that image quality assessment usually meets with the problem of perception-distortion (p-d) tradeoff. The existing deep image super-resolution (SR) methods either focus on high fidelity with pixel-level objectives or high perception with generative models. The emergence of diffusion model paves a fresh way for image restoration, which has the potential to offer a brand-new solution for p-d trade-off. We experimentally observed that the perceptual quality and distortion change in an opposite direction with the increase of sampling steps. In light of this property, we propose an adaptive skip diffusion model (SkipDiff), which aims to achieve high-fidelity perceptual image SR with fewer sampling steps. Specifically, it decouples the sampling procedure into coarse skip approximation and fine skip refinement stages. A coarse-grained skip diffusion is first performed as a high-fidelity prior to obtaining a latent approximation of the full diffusion. Then, a fine-grained skip diffusion is followed to further refine the latent sample for promoting perception, where the fine time steps are adaptively learned by deep reinforcement learning. Meanwhile, this approach also enables faster sampling of diffusion model through skipping the intermediate denoising process to shorten the effective steps of the computation. Extensive experimental results show that our SkipDiff achieves superior perceptual quality with plausible reconstruction accuracy and a faster sampling speed.

Efficient transformer-based models have made remarkable progress in image super-resolution (SR). Most of these works mainly design elaborate structures to accelerate the inference of the transformer, where all feature tokens are propagated equally. However, they ignore the underlying characteristic of image content, i.e., various image regions have distinct restoration difficulties, especially for large images (2K-8K), failing to achieve adaptive inference. In this work, we propose an adaptive token sparsification transformer (AdaFormer) to speed up the model inference for image SR. Specifically, a texture-relevant sparse attention block with parallel global and local branches is introduced, aiming to integrate informative tokens from the global view instead of only in fixed local windows. Then, an early-exit strategy is designed to progressively halt tokens according to the token importance. To estimate the plausibility of each token, we adopt a lightweight confidence estimator, which is constrained by an uncertainty-guided loss to obtain a binary halting mask about the tokens. Experiments on large images have illustrated that our proposal reduces nearly 90% latency against SwinIR on Test8K, while maintaining a comparable performance.

Federated learning (FL) provides a decentralized machine learning paradigm where a server collaborates with a group of clients to learn a global model without accessing the clients' data. User heterogeneity is a significant challenge for FL, which together with the class-distribution imbalance further enhances the difficulty of FL. Great progress has been made in large vision-language models, such as Contrastive Language-Image Pre-training (CLIP), which paves a new way for image classification and object recognition. Inspired by the success of CLIP on few-shot and zero-shot learning, we use CLIP to optimize the federated learning between server and client models under its vision-language supervision. It is promising to mitigate the user heterogeneity and class-distribution balance due to the powerful cross-modality representation and rich open-vocabulary prior knowledge. In this paper, we propose the CLIP-guided FL (CLIP2FL) method on heterogeneous and long-tailed data. In CLIP2FL, the knowledge of the off-the-shelf CLIP model is transferred to the client-server models, and a bridge is built between the client and server. Specifically, for client-side learning, knowledge distillation is conducted between client models and CLIP to improve the ability of client-side feature representation. For server-side learning, in order to mitigate the heterogeneity and class-distribution imbalance, we generate federated features to retrain the server model. A prototype contrastive learning with the supervision of the text encoder of CLIP is introduced to generate federated features depending on the client-side gradients, and they are used to retrain a balanced server classifier. Extensive experimental results on several benchmarks demonstrate that CLIP2FL achieves impressive performance and effectively deals with data heterogeneity and long-tail distribution. The code is available at https://github.com/shijiangming1/CLIP2FL.

Class-incremental object detection (CIOD) is a real-world desired capability, requiring an object detector to continuously adapt to new tasks without forgetting learned ones, with the main challenge being catastrophic forgetting. Many methods based on distillation and replay have been proposed to alleviate this problem. However, they typically learn on a pure visual backbone, neglecting the powerful representation capabilities of textual cues, which to some extent limits their performance. In this paper, we propose task-aware language-image representation to mitigate catastrophic forgetting, introducing a new paradigm for language-image-based CIOD. First of all, we demonstrate the significant advantage of language-image detectors in mitigating catastrophic forgetting. Secondly, we propose a learning task-aware language-image representation method that overcomes the existing drawback of directly utilizing the language-image detector for CIOD. More specifically, we learn the language-image representation of different tasks through an insulating approach in the training stage, while using the alignment scores produced by task-specific language-image representation in the inference stage. Through our proposed method, language-image detectors can be more practical for CIOD. We conduct extensive experiments on COCO 2017 and Pascal VOC 2007 and demonstrate that the proposed method achieves state-of-the-art results under the various CIOD settings.

Language conditioned 3D object grounding aims to find the object within the 3D scene mentioned by natural language descriptions, which mainly depends on the matching between visual and natural language. Considerable improvement in grounding performance is achieved by improving the multimodal fusion mechanism or bridging the gap between detection and matching. However, several mismatches are ignored, i.e., mismatch in local visual representation and global sentence representation, and mismatch in visual space and corresponding label word space. In this paper, we propose crossmodal match for 3D grounding from mitigating these mismatches perspective. Specifically, to match local visual features with the global description sentence, we propose BEV (Bird’s-eye-view) based global information embedding module. It projects multiple object proposal features into the BEV and the relations of different objects are accessed by the visual transformer which can model both positions and features with long-range dependencies. To circumvent the mismatch in feature spaces of different modalities, we propose crossmodal consistency learning. It performs cross-modal consistency constraints to convert the visual feature space into the label word feature space resulting in easier matching. Besides, we introduce label distillation loss and global distillation loss to drive these matches learning in a distillation way. We evaluate our method in mainstream evaluation settings on three datasets, and the results demonstrate the effectiveness of the proposed method.

Transformer has shown outstanding performance on image denoising, but the existing Transformer methods for image denoising are with large model sizes and high computational complexity, which is unfriendly to resource-constrained devices. In this paper, we propose a Lightweight Image Denoising Transformer method (LIDFormer) based on Triple Multi-Dconv Head Transposed Attention (TMDTA) to boost computational efficiency. LIDFormer first implements Discrete Wavelet Transform (DWT), which transforms the input image into a low-frequency space, greatly reducing the computational complexity of image denoising. However, the low-frequency image lacks fine-feature information, which degrades the denoising performance. To handle this problem, we introduce the Complementary Periodic Feature Reusing (CPFR) scheme for aggregating the shallow-layer features and the deep-layer features. Furthermore, TMDTA is proposed to integrate global context along three dimensions, thereby enhancing the ability of global feature representation. Note that our method can be applied as a pipeline for both convolutional neural networks and Transformers. Extensive experiments on several benchmarks demonstrate that the proposed LIDFormer achieves a better trade-off between high performance and low computational complexity on real-world image denoising tasks.

An effective pre-training framework with universal 3D representations is extremely desired in perceiving large- scale dynamic scenes. However, establishing such an ideal framework that is both task-generic and label-efficient poses a challenge in unifying the representation of the same primitive across diverse scenes. The current contrastive 3D pre-training methods typically follow a frame-level consistency, which focuses on the 2D-3D relationships in each detached image. Such inconsiderate consistency greatly hampers the promising path of reaching an universal pre-training framework: (1) The cross-scene semantic self-conflict, i.e., the intense collision between primitive segments of the same semantics from different scenes; (2) Lacking a globally unified bond that pushes the cross-scene semantic consistency into 3D representation learning. To address above challenges, we propose a CSC framework that puts a scene-level semantic consistency in the heart, bridging the connection of the similar semantic segments across various scenes. To achieve this goal, we combine the coherent semantic cues provided by the vision foundation model and the knowledge-rich cross-scene prototypes derived from the complementary multi-modality information. These allow us to train a universal 3D pre-training model that facilitates various downstream tasks with less fine-tuning efforts. Empirically, we achieve consistent improvements over SOTA pre-training approaches in semantic segmentation (+1.4% mIoU), object detection (+ 1.0% mAP), and panoptic segmentation (+3.0% PQ) using their task-specific 3D network on nuScenes. Code is released at https://github.com/chenhaomingbob/CSC, hoping to inspire future research.

The vision-language model has brought great improvement to few-shot industrial anomaly detection, which usually needs to design of hundreds of prompts through prompt engineering. For automated scenarios, we first use conventional prompt learning with many-class paradigm as the baseline to automatically learn prompts but found that it can not work well in one-class anomaly detection. To address the above problem, this paper proposes a one-class prompt learning method for few-shot anomaly detection, termed PromptAD. First, we propose semantic concatenation which can transpose normal prompts into anomaly prompts by concatenating normal prompts with anomaly suffixes, thus constructing a large number of negative samples used to guide prompt learning in one-class setting. Furthermore, to mitigate the training challenge caused by the absence of anomaly images, we introduce the concept of explicit anomaly margin, which is used to explicitly control the margin between normal prompt features and anomaly prompt features through a hyper-parameter. For image-level/pixel-level anomaly detection, PromptAD achieves first place in 11/12 few-shot settings on MVTec and VisA. Code is available at https://github.com/FuNz-0/PromptAD.git

The autonomous driving community has shown significant interest in 3D occupancy prediction, driven by its exceptional geometric perception and general object recognition capabilities. To achieve this, current works try to construct a Tri-Perspective View (TPV) or Occupancy (OCC) representation extending from the Bird-Eye-View perception. However, compressed views like TPV representation lose 3D geometry information while raw and sparse OCC representation requires heavy but redundant computational costs. To address the above limitations, we propose Compact Occupancy TRansformer (COTR), with a geometry-aware occupancy encoder and a semantic-aware group decoder to reconstruct a compact 3D OCC representation. The occupancy encoder first generates a compact geometrical OCC feature through efficient explicit-implicit view transformation. Then, the occupancy decoder further enhances the semantic discriminability of the compact OCC representation by a coarse-to-fine semantic grouping strategy. Empirical experiments show that there are evident performance gains across multiple baselines, e.g., COTR outperforms baselines with a relative improvement of 8%-15%, demonstrating the superiority of our method. The code is available at https://github.com/NotACracker/COTR.

Scene text recognition (STR) in the wild frequently en-counters challenges when coping with domain variations, font diversity, shape deformations, etc. A straightforward solution is performing model fine-tuning tailored to a spe-cific scenario, but it is computationally intensive and re-quires multiple model copies for various scenarios. Re-cent studies indicate that large language models (LLMs) can learn from afew demonstration examples in a training-free manner, termed “In-Context Learning” (ICL). Never-theless, applying LLMs as a text recognizer is unacceptably resource-consuming. Moreover, our pilot experiments on LLMs show that ICL fails in STR, mainly attributed to the insufficient incorporation of contextual information from di-verse samples in the training stage. To this end, we intro-duce E2 STR, a STR model trained with context-rich scene text sequences, where the sequences are generated via our proposed in-context training strategy. E2 STR demonstrates that a regular-sized model is sufficient to achieve effective ICL capabilities in STR. Extensive experiments show that E2 STR exhibits remarkable training-free adaptation in var-ious scenarios and outperforms even the fine-tuned state-of-the-art approaches on public benchmarks. The code is released at https://github.com/bytedanceIE2STR.

Prompt-tuning has demonstrated impressive performance in continual learning by querying relevant prompts for each input instance, which can avoid the introduction of task identifier. Its forgetting is therefore reduced as this instance-wise query mechanism enables us to select and update only relevant prompts. In this paper, we further integrate prompt-tuning with gradient projection approach. Our observation is: prompt-tuning releases the necessity of task identifier for gradient projection method; and gradient projection provides theoretical guarantees against forgetting for prompt-tuning. This inspires a new prompt gradient projection approach (PGP) for continual learning. In PGP, we deduce that reaching the orthogonal condition for prompt gradient can effectively prevent forgetting via the self-attention mechanism in vision-transformer. The condition equations are then realized by conducting Singular Value Decomposition (SVD) on an element-wise sum space between input space and prompt space. We validate our method on diverse datasets and experiments demonstrate the efficiency of reducing forgetting both in class incremental, online class incremental, and task incremental settings. The code is available at https://github.com/JingyangQiao/prompt-gradient-projection.

Point cloud semantic segmentation heavily relies on the large-scale point-level annotated dataset, which encourages the weakly-supervised method to prevail gradually. Previous weakly-supervised self-training methods only adopted positive labels, which would be under-performed due to too much noise and lack of supervision. We are the first to present negative labels into the 3D segmentation area, providing extra supervision for hard samples to mitigate the drawbacks induced by noisy labels. Together with the positive labels, we formulate a Mutual Positive-Negative Bi-branch Learning framework to generate positive and negative labels iteratively. Based on that positive branch and negative branch learn complementary knowledge, we build a Mutual Positive-Negative Knowledge Distillation within the bi-branch to further encourage the two branches to learn from each other. Finally, we propose a novel dynamic fusion strategy to fuse predictions from the positive and negative branches, generating more robust predictions. Results on three large-scale datasets show that our method outperforms state-of-the-art weakly-supervised methods by a large margin.

Deep neural networks have revealed enormous potential in video super-resolution (VSR), yet the expensive computational expense limits their deployment on resource-limited devices and actual scenarios, especially for restoring multiple frames simultaneously. Existing VSR models contain considerable redundant filters, which drag down the inference efficiency. To accelerate the inference of VSR models, we propose a scalable method based on adaptive patch routing to achieve practical speedup. Specifically, we design a confidence estimator to predict the aggregation performance of each block for adjacent patch information. It learns to dynamically perform block skipping, i.e., choose which basic blocks of the VSR network to execute during inference so as to reduce total computation to the maximum extent without degrading reconstruction accuracy dramatically. However, we observe that skipping error would be amplified as the hidden states propagate along with recurrent networks. To alleviate the issue, we design temporal feature alignment to guarantee the performance. This proposal essentially proposes an adaptive routing scheme for each patch. Extensive experiments demonstrate that our method can not only accelerate inference but also provide strong quantitative and qualitative results. Built upon the BasicVSR model, our method achieves a speedup of 20% on average, going as high as 50% for some images, while even maintaining competitive performance on REDS4.

Multi-modal Unsupervised Domain Adaptation (MM-UDA) for large-scale 3D semantic segmentation involves adapting 2D and 3D models to a target domain without labels, which significantly reduces the labor-intensive annotations. Existing MM-UDA methods have often attempted to mitigate the domain discrepancy by aligning features between the source and target data. However, this implementation falls short when applied to image perception due to the susceptibility of images to environmental changes compared to point clouds. To mitigate this limitation, in this work, we explore the potentials of an off-the-shelf Contrastive Language-Image Pre-training (CLIP) model with rich whilst heterogeneous knowledge. To make CLIP task-specific, we propose a top-performing method, dubbed CLIP2UDA, which makes frozen CLIP reward unsupervised domain adaptation in 3D semantic segmentation. Specifically, CLIP2UDA alternates between two steps during adaptation: (a) Learning task-specific prompt. 2D features response from the visual encoder are employed to initiate the learning of adaptive text prompt of each domain, and (b) Learning multi-modal domain-invariant representations. These representations interact hierarchically in the shared decoder to obtain unified 2D visual predictions. This enhancement allows for effective alignment between the modality-specific 3D and unified feature space via cross-modal mutual learning. Extensive experimental results demonstrate that our method outperforms state-of-the-art competitors in several widely-recognized adaptation scenarios. Code is available at: https://github.com/Barcaaaa/CLIP2UDA.

Unsupervised Visible-Infrared Person Re-identification (USVI-ReID) presents a formidable challenge, which aims to match pedestrian images across visible and infrared modalities without any annotations. Recently, clustered pseudo-label methods have become predominant in USVI-ReID, although the inherent noise in pseudo-labels presents a significant obstacle. Most existing works primarily focus on shielding the model from the harmful effects of noise, neglecting to calibrate noisy pseudo-labels usually associated with hard samples, which will compromise the robustness of the model. To address this issue, we design a Robust Pseudo-label Learning with Neighbor Relation (RPNR) framework for USVI-ReID. To be specific, we first introduce a straightforward yet potent Noisy Pseudo-label Calibration module to correct noisy pseudo-labels. Due to the high intra-class variations, noisy pseudo-labels are difficult to calibrate completely. Therefore, we introduce a Neighbor Relation Learning module to reduce high intra-class variations by modeling potential interactions between all samples. Subsequently, we devise an Optimal Transport Prototype Matching module to establish reliable cross-modality correspondences. On that basis, we design a Memory Hybrid Learning module to jointly learn modality-specific and modality-invariant information. Comprehensive experiments conducted on two widely recognized benchmarks, SYSU-MM01 and RegDB, demonstrate that RPNR outperforms the current state-of-the-art GUR with an average Rank-1 improvement of 10.3%. The code is available at https://github.com/XiangboYin/RPNR.

Open-vocabulary semantic segmentation (OVSS) aims to segment unseen classes without corresponding labels. Existing Vision-Language Model (VLM)-based methods leverage VLM's rich knowledge to enhance additional explicit segmentation-specific networks, yielding competitive results, but at the cost of extensive training cost. To reduce the cost, we attempt to enable VLM to directly produce the segmentation results without any segmentation-specific networks. Prompt learning offers a direct and parameter-efficient approach, yet it falls short in guiding VLM for pixel-level visual classification. Therefore, we propose the ${\bf R}$elationship ${\bf P}$rompt ${\bf M}$odule (${\bf RPM}$), which generates the relationship prompt that directs VLM to extract pixel-level semantic embeddings suitable for OVSS. Moreover, RPM integrates with VLM to construct the ${\bf R}$elationship ${\bf P}$rompt ${\bf N}$etwork (${\bf RPN}$), achieving OVSS without any segmentation-specific networks. RPN attains state-of-the-art performance with merely about ${\bf 3M}$ trainable parameters (2\% of total parameters).

3D semantic segmentation using an adapting model trained from a source domain with or without accessing unlabeled target-domain data is the fundamental task in computer vision, containing domain adaptation and domain generalization. The essence of simultaneously solving cross-domain tasks is to enhance the generalizability of the encoder. In light of this, we propose a groundbreaking universal method with the help of off-the-shelf Visual Foundation Models (VFMs) to boost the adaptability and generalizability of cross-domain 3D semantic segmentation, dubbed $\textbf{UniDSeg}$. Our method explores the VFMs prior and how to harness them, aiming to inherit the recognition ability of VFMs. Specifically, this method introduces layer-wise learnable blocks to the VFMs, which hinges on alternately learning two representations during training: (i) Learning visual prompt. The 3D-to-2D transitional prior and task-shared knowledge is captured from the prompt space, and then (ii) Learning deep query. Spatial Tunability is constructed to the representation of distinct instances driven by prompts in the query space. Integrating these representations into a cross-modal learning framework, UniDSeg efficiently mitigates the domain gap between 2D and 3D modalities, achieving unified cross-domain 3D semantic segmentation. Extensive experiments demonstrate the effectiveness of our method across widely recognized tasks and datasets, all achieving superior performance over state-of-the-art methods. Remarkably, UniDSeg achieves 57.5\%/54.4\% mIoU on ``A2D2/sKITTI'' for domain adaptive/generalized tasks. Code is available at https://github.com/Barcaaaa/UniDSeg.

In this work, we present TextHarmony, a unified and versatile multimodal generative model proficient in comprehending and generating visual text. Simultaneously generating images and texts typically results in performance degradation due to the inherent inconsistency between vision and language modalities. To overcome this challenge, existing approaches resort to modality-specific data for supervised fine-tuning, necessitating distinct model instances. We propose Slide-LoRA, which dynamically aggregates modality-specific and modality-agnostic LoRA experts, partially decoupling the multimodal generation space. Slide-LoRA harmonizes the generation of vision and language within a singular model instance, thereby facilitating a more unified generative process. Additionally, we develop a high-quality image caption dataset, DetailedTextCaps-100K, synthesized with a sophisticated closed-source MLLM to enhance visual text generation capabilities further. Comprehensive experiments across various benchmarks demonstrate the effectiveness of the proposed approach. Empowered by Slide-LoRA, TextHarmony achieves comparable performance to modality-specific fine-tuning results with only a 2% increase in parameters and shows an average improvement of 2.5% in visual text comprehension tasks and 4.0% in visual text generation tasks. Our work delineates the viability of an integrated approach to multimodal generation within the visual text domain, setting a foundation for subsequent inquiries. Code is available at https://github.com/bytedance/TextHarmony.

Combining LiDAR points and images for robust semantic segmentation has shown great potential. However, the heterogeneity between the two modalities (e.g. the density, the field of view) poses challenges in establishing a bijective mapping between each point and pixel. This modality alignment problem introduces new challenges in network design and data processing for cross-modal methods. Specifically, 1) points that are projected outside the image planes; 2) the complexity of maintaining geometric consistency limits the deployment of many data augmentation techniques. To address these challenges, we propose a cross-modal knowledge imputation and transition approach. First, we introduce a bidirectional feature fusion strategy that imputes missing image features and performs cross-modal fusion simultaneously. This allows us to generate reliable predictions even when images are missing. Second, we propose a Uni-to-Multi modal Knowledge Distillation (U2MKD) framework, leveraging the transfer of informative features from a single-modality teacher to a cross-modality student. This overcomes the issues of augmentation misalignment and enables us to train the student effectively. Extensive experiments on the nuScenes, Waymo, and SemanticKITTI datasets demonstrate the effectiveness of our approach. Notably, our method achieves an 8.3 mIoU gain over the LiDAR-only baseline on the nuScenes validation set and achieves state-of-the-art performance on the three datasets.

Information Bottleneck (IB) provides an information-theoretic principle for multi-view learning by revealing the various components contained in each viewpoint. This highlights the necessity to capture their distinct roles to achieve view-invariance and predictive representations but remains under-explored due to the technical intractability of modeling and organizing innumerable mutual information (MI) terms. Recent studies show that sufficiency and consistency play such key roles in multi-view representation learning, and could be preserved via a variational distillation framework. But when it generalizes to arbitrary viewpoints, such strategy fails as the mutual information terms of consistency become complicated. This paper presents Multi-View Variational Distillation (MV$^{2}$D), tackling the above limitations for generalized multi-view learning. Uniquely, MV$^{2}$D can recognize useful consistent information and prioritize diverse components by their generalization ability. This guides an analytical and scalable solution to achieving both sufficiency and consistency. Additionally, by rigorously reformulating the IB objective, MV$^{2}$D tackles the difficulties in MI optimization and fully realizes the theoretical advantages of the information bottleneck principle. We extensively evaluate our model on diverse tasks to verify its effectiveness, where the considerable gains provide key insights into achieving generalized multi-view representations under a rigorous information-theoretic principle.

Glass surface detection is challenging as glass normally borrows similar visual appearances from the arbitrary objects/scenes behind it. Although some methods have been proposed to address this problem, they may fail if the reference objects are nonexistent or the additional annotations are missing. This article aims to address the glass surface detection problem by utilizing the intrinsic glass properties without reference objects and additional annotations. We observe glass makes blurs naturally. Based on the investigation of this intrinsic visual blurriness cue, we propose a novel visual blurriness aggregation module to model visual blurriness as a learnable residual in order to extract and aggregate multiscale valuable visual blurriness features used for guiding the backbone features to detect glass precisely. Besides, we note the ratio of the blurred area assists in utilizing the visual blurriness cue caused by glass and propose a visual blurriness driven refinement module to refine glass maps with this ratio to better leverage the visual blurriness information. Extensive experiments show that the proposed method achieves state-of-the-art performance on popular glass surface datasets.

Weakly supervised point cloud semantic segmentation methods that require 1% or fewer labels with the aim of realizing almost the same performance as fully supervised approaches have recently attracted extensive research attention. A typical solution in this framework is to use self-training or pseudo-labeling to mine the supervision from the point cloud itself while ignoring the critical information from images. In fact, cameras widely exist in LiDAR scenarios, and this complementary information seems to be highly important for 3D applications. In this paper, we propose a novel cross-modality weakly supervised method for 3D segmentation that incorporates complementary information from unlabeled images. We design a dual-branch network equipped with an active labeling strategy to maximize the power of tiny parts of labels and to directly realize 2D-to-3D knowledge transfer. Afterward, we establish a cross-modal self-training framework, which iterates between parameter updating and pseudolabel estimation. In the training phase, we propose cross-modal association learning to mine complementary supervision from images by reinforcing the cycle consistency between 3D points and 2D superpixels. In the pseudolabel estimation phase, a pseudolabel self-rectification mechanism is derived to filter noisy labels, thus providing more accurate labels for the networks to be fully trained. The extensive experimental results demonstrate that our method even outperforms the state-of-the-art fully supervised competitors with less than 1% actively selected annotations.

Night-time scene parsing aims to extract pixel-level semantic information in night images, aiding downstream tasks in understanding scene object distribution. Due to limited labeled night image datasets, unsupervised domain adaptation (UDA) has become the predominant method for studying night scenes. UDA typically relies on paired day-night image pairs to guide adaptation, but this approach hampers dataset construction and restricts generalization across night scenes in different datasets. Moreover, UDA, focusing on network architecture and training strategies, faces difficulties in handling classes with few domain similarities. In this paper, we leverage Prompt Images Guidance (PIG) to enhance UDA with supplementary night knowledge. We propose a Night-Focused Network (NFNet) to learn night-specific features from both target domain images and prompt images. To generate high-quality pseudo-labels, we propose Pseudo-label Fusion via Domain Similarity Guidance (FDSG). Classes with fewer domain similarities are predicted by NFNet, which excels in parsing night features, while classes with more domain similarities are predicted by UDA, which has rich labeled semantics. Additionally, we propose two data augmentation strategies: the Prompt Mixture Strategy (PMS) and the Alternate Mask Strategy (AMS), aimed at mitigating the overfitting of the NFNet to a few prompt images. We conduct extensive experiments on four night-time datasets: NightCity, NightCity+, Dark Zurich, and ACDC. The results indicate that utilizing PIG can enhance the parsing accuracy of UDA. The code is available at https://github.com/qiurui4shu/PIG.

Mirror detection is a challenging task since mirrors do not possess a consistent visual appearance. Even the Segment Anything Model (SAM), which boasts superior zero-shot performance, cannot accurately detect the position of mirrors. Existing methods determine the position of the mirror under hypothetical conditions, such as the correspondence between objects inside and outside the mirror, and the semantic association between the mirror and surrounding objects. However, these assumptions do not apply to all scenarios. For instance, there may be no corresponding real objects to the reflected objects in the scene, or it may be challenging to extract meaningful semantic associations in complex scenes. On the other hand, humans can easily recognize mirrors through the specular texture caused by materials. To mine mirror features in more general scenes, we propose a Cross-Space-Frequency Window Transformer (CSFwinformer) to extract spatial and frequency features for texture analysis. Specifically, we design a Spatial-Frequency Window Alignment module (SFWA) to calculate spatial-frequency feature affinities and learn the difference between mirror and non-mirror textures. We then propose a Dilated Window Attention (DWA) to extract global features to complement the limitation of window alignment. Besides, we propose a Cross-Modality Context Contrast module (CMCC) to fuse cross-modality features and global features, which enables information flow between different windows to take full advantage of cross-modality information. Extensive experiments show that our method performs favorably against state-of-the-art methods on three mirror detection benchmarks and significantly improved SAM performance on mirror detection. The code is available at https://github.com/wangsen99/CSFwinformer.

Multiview clustering has become a research hotspot in recent years due to its excellent capability of heterogeneous data fusion. Although a great deal of related works has appeared one after another, most of them generally overlook the potentials of prior knowledge utilization and progressive sample learning, resulting in unsatisfactory clustering performance in real-world applications. To deal with the aforementioned drawbacks, in this article, we propose a semisupervised progressive representation learning approach for deep multiview clustering (namely, SPDMC). Specifically, to make full use of the discriminative information contained in prior knowledge, we design a flexible and unified regularization, which models the sample pairwise relationship by enforcing the learned view-specific representation of must-link (ML) samples (cannot-link (CL) samples) to be similar (dissimilar) with cosine similarity. Moreover, we introduce the self-paced learning (SPL) paradigm and take good care of two characteristics in terms of both complexity and diversity when progressively learning multiview representations, such that the complementarity across multiple views can be squeezed thoroughly. Through comprehensive experiments on eight widely used image datasets, we prove that the proposed approach can perform better than the state-of-the-art opponents.

Recently, with the development of intelligent manufacturing, the demand for surface defect inspection has been increasing. Deep learning has achieved promising results in defect inspection. However, due to the rareness of defect data and the difficulties of pixelwise annotation, the existing supervised defect inspection methods are too inferior to be implemented in practice. To solve the problem of defect segmentation with few labeled data, we propose a simple and efficient method for semisupervised defect segmentation (SSDS), named perturbed progressive learning (PPL). On the one hand, PPL decouples the predictions of student and teacher networks as well as alleviates overfitting on noisy pseudo-labels. On the other hand, PPL encourages consistency across various perturbations in a broader stagewise scope, alleviating drift caused by the noisy pseudo-labels. Specifically, PPL contains two training stages. In the first stage, the teacher network gives the unlabeled data with pseudo-labels that are divided into the easy and hard groups. The labeled data and the unlabeled data in the easy group with their perturbation are both used to train for a better-performing student network. In the second stage, the unlabeled data in the hard group are predicted by the obtained student network, so the refined pseudo-labeled data are enlarged. All the pseudo-labeling data and labeled data with their perturbation are used to retrain the student network, progressively improving the defect feature representation. We build a mobile screen defect dataset (MSDD-3) with three classes of defects. PPL is implemented on MSDD-3 as well as other public datasets. Extensive experimental results demonstrate that PPL significantly surpasses the state-of-the-art methods across all evaluation partition protocols.

Multiview clustering via binary representation has attracted intensive attention due to its effectiveness in handling large-scale multiple view data. However, these kind of clustering approaches usually ignore a very important potential high-order correlation in discrete representation learning. In this article, we propose a novel all-in collaborative multiview binary representation for clustering (AC-MVBC) framework, where multiview collaborative binary representation and clustering structure are learned in a joint manner. Specifically, using a new type of tensor low-rank constraint, the high-order collaborations, i.e., cross-view and inner view collaborations, can be effectively captured in our model. Moreover, by incorporating the Bregman discrepancy, the projective consistency among different views can be guaranteed to achieve a more powerful binary representation. An efficient optimization algorithm is also proposed to solve the objective function with fast convergence empirically. Experimental results on several challenge datasets demonstrate that the proposed method has achieved highly competent performance compared with the state-of-the-art multiview clustering (MVC) methods while maintaining low computational and memory requirements.

Recent Few-Shot Learning (FSL) methods put emphasis on generating a discriminative embedding features to precisely measure the similarity between support and query sets. Current CNN-based cross-attention approaches generate discriminative representations via enhancing the mutually semantic similar regions of support and query pairs. However, it suffers from two problems: CNN structure produces inaccurate attention map based on local features, and mutually similar backgrounds cause distraction. To alleviate these problems, we design a novel SpatialFormer structure to generate more accurate attention regions based on global features. Different from the traditional Transformer modeling intrinsic instance-level similarity which causes accuracy degradation in FSL, our SpatialFormer explores the semantic-level similarity between pair inputs to boost the performance. Then we derive two specific attention modules, named SpatialFormer Semantic Attention (SFSA) and SpatialFormer Target Attention (SFTA), to enhance the target object regions while reduce the background distraction. Particularly, SFSA highlights the regions with same semantic information between pair features, and SFTA finds potential foreground object regions of novel feature that are similar to base categories. Extensive experiments show that our methods are effective and achieve new state-of-the-art results on few-shot classification benchmarks.

As manual point-wise label is time and labor-intensive for fully supervised large-scale point cloud semantic segmentation, weakly supervised method is increasingly active. However, existing methods fail to generate high-quality pseudo labels effectively, leading to unsatisfactory results. In this paper, we propose a weakly supervised point cloud semantic segmentation framework via receptive-driven pseudo label consistency and structural consistency to mine potential knowledge. Specifically, we propose three consistency contrains: pseudo label consistency among different scales, semantic structure consistency between intra-class features and class-level relation structure consistency between pair-wise categories. Three consistency constraints are jointly used to effectively prepares and utilizes pseudo labels simultaneously for stable training. Finally, extensive experimental results on three challenging datasets demonstrate that our method significantly outperforms state-of-the-art weakly supervised methods and even achieves comparable performance to the fully supervised methods.

The last decades are marked by massive and diverse image data, which shows increasingly high resolution and quality. However, some images we obtained may be corrupted, affecting the perception and the application of downstream tasks. A generic method for generating a high-quality image from the degraded one is in demand. In this paper, we present a novel GAN inversion framework that utilizes the powerful generative ability of StyleGAN-XL for this problem. To ease the inversion challenge with StyleGAN-XL, Clustering \& Regularize Inversion (CRI) is proposed. Specifically, the latent space is firstly divided into finer-grained sub-spaces by clustering. Instead of initializing the inversion with the average latent vector, we approximate a centroid latent vector from the clusters, which generates an image close to the input image. Then, an offset with a regularization term is introduced to keep the inverted latent vector within a certain range. We validate our CRI scheme on multiple restoration tasks (i.e., inpainting, colorization, and super-resolution) of complex natural images, and show preferable quantitative and qualitative results. We further demonstrate our technique is robust in terms of data and different GAN models. To our best knowledge, we are the first to adopt StyleGAN-XL for generating high-quality natural images from diverse degraded inputs. Code is available at https://github.com/Booooooooooo/CRI.

Incremental learning could be roughly divided into two categories, i.e., class- and task-incremental learning. The main difference is whether the task ID is given during evaluation. In this paper, we show this task information is indeed a strong prior knowledge, which will bring significant improvement over class-incremental learning baseline, e.g., DER [39]. Based on this observation, we propose a gate network to predict the task ID for class incremental inference. This is challenging as there is no explicit semantic relationship between categories in the concept of task. Therefore, we propose a multi-centroid task descriptor by assuming the data within a task can form multiple clusters. The cluster centers are optimized by pulling relevant sample-centroid pairs while pushing others away, which ensures that there is at least one centroid close to a given sample. To select relevant pairs, we use class prototypes as proxies and solve a bipartite matching problem, making the task descriptor representative yet not degenerate to uni-modal. As a result, our dynamic inference network is trained independently of baseline and provides a flexible, efficient solution to distinguish between tasks. Extensive experiments show our approach achieves state-of-the-art results, e.g., we achieve 72.41% average accuracy on CIFAR100-BOS50, outperforming DER by 3.40%.

Scalable deep Super-Resolution (SR) models are increasingly in demand, whose memory can be customized and tuned to the computational recourse of the platform. The existing dynamic scalable SR methods are not memory-friendly enough because multi-scale models have to be saved with a fixed size for each model. Inspired by the success of Lottery Tickets Hypothesis (LTH) on image classification, we explore the existence of unstructured scalable SR deep models, that is, we find gradual shrinkage subnetworks of extreme sparsity named winning tickets. In this paper, we propose a Memory-friendly Scalable SR framework (MSSR). The advantage is that only a single scalable model covers multiple SR models with different sizes, instead of reloading SR models of different sizes. Concretely, MSSR consists of the forward and backward stages, the former for model compression and the latter for model expansion. In the forward stage, we take advantage of LTH with rewinding weights to progressively shrink the SR model and the pruning-out masks that form nested sets. Moreover, stochastic self-distillation (SSD) is conducted to boost the performance of sub-networks. By stochastically selecting multiple depths, the current model inputs the selected features into the corresponding parts in the larger model and improves the performance of the current model based on the feedback results of the larger model. In the backward stage, the smaller SR model could be expanded by recovering and fine-tuning the pruned parameters according to the pruning-out masks obtained in the forward. Extensive experiments show the effectiveness of MMSR. The smallest-scale sub-network could achieve the sparsity of 94% and outperforms the compared lightweight SR methods.

Continual learning aims to incrementally learn novel classes over time, while not forgetting the learned knowledge. Recent studies have found that learning would not forget if the updated gradient is orthogonal to the feature space. However, previous approaches require the gradient to be fully orthogonal to the whole feature space, leading to poor plasticity, as the feasible gradient direction becomes narrow when the tasks continually come, i.e., feature space is unlimitedly expanded. In this paper, we propose a space decoupling (SD) algorithm to decouple the feature space into a pair of complementary subspaces, i.e., the stability space$\mathcal{I}$and the plasticity space$\mathcal{R}. \mathcal{I}$is established by conducting space intersection between the historic and current feature space, and thus$\mathcal{I}$contains more task-shared bases.$\mathcal{R}$is constructed by seeking the orthogonal complementary subspace of$T$and thus$\mathcal{R}$mainly contains task-specific bases. By putting distinguishing constraints on$\mathcal{R}$and$\mathcal{I}$, our method achieves a better balance between stability and plasticity. Extensive experiments are conducted by applying SD to gradient projection baselines, and show SD is model-agnostic and achieves SOTA results on publicly available datasets.

Visible-infrared person re-identification (VI-ReID) aims to match a specific person from a gallery of images captured from non-overlapping visible and infrared cameras. Most works focus on fully supervised VI-ReID, which requires substantial cross-modality annotation that is more expensive than the annotation in single-modality. To reduce the extensive cost of annotation, we explore two practical semi-supervised settings: uni-semi-supervised (annotating only visible images) and bi-semi-supervised (annotating partially in both modalities). These two semi-supervised settings face two challenges due to the large cross-modality discrepancies and the lack of correspondence supervision between visible and infrared images. Thus, it is diffi-cult to generate reliable pseudo-labels and learn modality-invariant features from noise pseudo-labels. In this paper, we propose a dual pseudo-label interactive self-training (DPIS) for these two semi-supervised VI-ReID. Our DPIS integrates two pseudo-labels generated by distinct models into a hybrid pseudo-label for unlabeled data. However, the hybrid pseudo-label still inevitably contains noise. To eliminate the negative effect of noise pseudo-labels, we introduce three modules: noise label penalty (NLP), noise correspondence calibration (NCC), and unreliable anchor learning (UAL). Specifically, NLP penalizes noise labels, NCC calibrates noisy correspondences, and UAL mines the hard-to-discriminate features. Extensive experimental results on SYSU-MM01 and RegDB demonstrate that our DPIS achieves impressive performance under these two semi-supervised settings.

Continual Learning (CL) is the constant development of complex behaviors by building upon previously acquired skills. Yet, current CL algorithms tend to incur class-level forgetting as the label information is often quickly overwritten by new knowledge. This motivates attempts to mine instance-level discrimination by resorting to recent self-supervised learning (SSL) techniques. However, previous works have pointed out that the self-supervised learning objective is essentially a trade-off between invariance to distortion and preserving sample information, which seriously hinders the unleashing of instance-level discrimination.In this work, we reformulate SSL from the information-theoretic perspective by disentangling the goal of instance-level discrimination, and tackle the trade-off to promote compact representations with maximally preserved invariance to distortion. On this basis, we develop a novel alternate learning paradigm to enjoy the complementary merits of instance-level and category-level supervision, which yields improved robustness against forgetting and better adaptation to each task. To verify the proposed method, we conduct extensive experiments on four different benchmarks using both class-incremental and task-incremental settings, where the leap in performance and thorough ablation studies demonstrate the efficacy and efficiency of our modeling strategy.

3D panoptic segmentation is a challenging perception task that requires both semantic segmentation and instance segmentation. In this task, we notice that images could provide rich texture, color, and discriminative information, which can complement LiDAR data for evident performance improvement, but their fusion remains a challenging problem. To this end, we propose LCPS, the first LiDAR-Camera Panoptic Segmentation network. In our approach, we conduct LiDAR-Camera fusion in three stages: 1) an Asynchronous Compensation Pixel Alignment (ACPA) module that calibrates the coordinate misalignment caused by asynchronous problems between sensors; 2) a Semantic-Aware Region Alignment (SARA) module that extends the one-to-one point-pixel mapping to one-to-many semantic relations; 3) a Point-to-Voxel feature Propagation (PVP) module that integrates both geometric and semantic fusion information for the entire point cloud. Our fusion strategy improves about 6.9% PQ performance over the LiDAR-only baseline on NuScenes dataset. Extensive quantitative and qualitative experiments further demonstrate the effectiveness of our novel framework. The code will be released at https://github.com/zhangzw12319/lcps.git.

Unlike conventional zero-shot learning (CZSL) which only focuses on the recognition of unseen classes by using the classifier trained on seen classes and semantic embeddings, generalized zero-shot learning (GZSL) aims at recognizing both the seen and unseen classes, so it is more challenging due to the extreme training imbalance. Recently, some feature generation methods introduce metric learning to enhance the discriminability of visual features. Although these methods achieve good results, they focus only on metric learning in the visual feature space to enhance features and ignore the association between the feature space and the semantic space. Since the GZSL method uses semantics as prior knowledge to migrate visual knowledge to unseen classes, the consistency between visual space and semantic space is critical. To this end, we propose relational metric learning which can relate the metrics in the two spaces and make the distribution of the two spaces more consistent. Based on the generation method and relational metric learning, we proposed a novel GZSL method, termed VS-Boost, which can effectively boost the association between vision and semantics. The experimental results demonstrate that our method is effective and achieves significant gains on five benchmark datasets compared with the state-of-the-art methods.

Large-scale Vision-Language Pre-training (VLP) model, e.g., CLIP, has demonstrated its natural advantage in generating textual descriptions for images. These textual descriptions afford us greater semantic monitoring insights while not requiring any domain knowledge. In this paper, we propose a new prompt learning paradigm for unsupervised visible-infrared person re-identification (USL-VI-ReID) by taking full advantage of the visual-text representation ability from CLIP. In our framework, we establish a learnable cluster-aware prompt for person images and obtain textual descriptions allowing for subsequent unsupervised training. This description complements the rigid pseudo-labels and provides an important semantic supervised signal. On that basis, we propose a new memory-swapping contrastive learning, where we first find the correlated cross-modal prototypes by the Hungarian matching method and then swap the prototype pairs in the memory. Thus typical contrastive learning without any change could easily associate the cross-modal information. Extensive experiments on the benchmark datasets demonstrate the effectiveness of our method. For example, on SYSU-MM01 we arrive at 54.0% in terms of Rank-1 accuracy, over 9% improvement against state-of-the-art approaches. Code is available at https://github.com/CzAngus/CCLNet.

Cross-modal recipe retrieval is an emerging visual-textual retrieval task, which aims at matching food images with the corresponding recipes. Although large-scale Vision-Language Pre-training (VLP) models have achieved impressive performance on a wide range of downstream tasks, they still perform unsatisfactorily on this cross-modal retrieval task due to the following two problems: (1) Features from food images and recipes need to be aligned, simply fine-tuning the pre-trained VLP model's image encoder does not explicitly help with this goal. (2) The text content in the recipe is more structured than the text caption in the VLP model's pre-training corpus, which prevents the VLP model from adapting to the recipe retrieval task. In this paper, we propose a Component-aware Instance-specific Prompt learning (CIP) model that fully exploits the ability of large-scale VLP models. CIP enables us to learn the structured recipe information and therefore allows for aligning visual-textual representations without fine-tuning. Furthermore, we construct a recipe encoder termed Adaptive Recipe Merger (ARM) based on hierarchical Transformers, encouraging the model to learn more effective recipe representations. Extensive experiments on the public Recipe1M dataset demonstrate the superiority of our proposed method by outperforming the state-of-the-art methods on cross-modal recipe retrieval task.

Cross-modal Unsupervised Domain Adaptation (UDA) becomes a research hotspot because it reduces the laborious annotation of target domain samples. Existing methods only mutually mimic the outputs of cross-modality in each domain, which enforces the class probability distribution agreeable in different domains. However, these methods ignore the complementarity brought by the modality fusion representation in cross-modal learning. In this paper, we propose a cross-modal UDA method for 3D semantic segmentation via Bidirectional Fusion-then-Distillation, named BFtD-xMUDA, which explores cross-modal fusion in UDA and realizes distribution consistency between outputs of two domains not only for 2D image and 3D point cloud but also for 2D/3D and fusion. Our method contains three significant components: Model-agnostic Feature Fusion Module (MFFM), Bidirectional Distillation (B-Distill), and Cross-modal Debiased Pseudo-Labeling (xDPL). MFFM is employed to generate cross-modal fusion features for establishing a latent space, which enforces maximum correlation and complementarity between two heterogeneous modalities. B-Distill is introduced to exploit bidirectional knowledge distillation which includes cross-modality and cross-domain fusion distillation, and well-achieving domain-modality alignment. xDPL is designed to model the uncertainty of pseudo-labels by self-training scheme. Extensive experimental results demonstrate that our method outperforms state-of-the-art competitors in several adaptation scenarios.

Single image super-resolution (SR) is an important low-level vision task, and the dynamic SR trading off performance and efficiency are increasingly in demand. The existing dynamic SR methods are divided into two classes: the structured pruning and non-structured compressing methods. The former removes redundant structures in the network, which often leads to significant performance degradation, and the latter searches for extremely sparse parameter masks, achieving promising performance, but they are not deployable in hardware platforms with irregular memory access. In order to solve the mentioned problems, we propose Hardware-friendly Scalable SR (HSSR) with progressively structured sparsity. The superiority of our method is that with only a single scalable model it covers multiple SR models with different sizes, without extra retraining or post-processing. HSSR contains the forward and backward processing. In the forward process, we gradually shrink the SR networks with structured iterative sparsity where grouping convolution together with knowledge distillation is conducted to reduce the amount of SR parameters and the computational complexity while keeping the performance, and in the backward process, we gradually expand the compressed SR networks with structured iterative recovery. Comprehensive experiments on benchmark datasets show that HSSR is perfectly compatible with common convolution baselines. Compared with the Slimmable method, our model is superior in performance, flops, and model size. Experimental results demonstrate that HSSR achieves significant compression, saving up to 1500K parameters and 100 GFlops calculation compared to the original model in real-world applications.

Single image super-resolution (SISR) has made a significant breakthrough benefiting from the prevalent rise of deep neural networks and large-scale training samples. The mainstream deep SR models primarily focus on network architecture design as well as optimization schemes, while few pay attention to the training data. In fact, most of the existing SR methods train the model on uniformly sampled patch pairs from the whole image. However, the uneven image content makes the training data present an unbalanced distribution, i.e., the easily reconstructed region (smooth) occupies the majority of the data, while the hard reconstructed region (edge or texture) has rarely few samples. Based on this phenomenon, we consider rethinking the current paradigm of merely using uniform data sampling way for training SR models. In this paper, we propose a simple yet effective Bi-Sampling Parameter Attribution (BSPA) method for accurate image SR. Specifically, the bi-sampling consists of uniform sampling and inverse sampling, which is introduced to reconcile the unbalanced inherent data bias. The former aims to keep the intrinsic data distribution, and the latter is designed to enhance the feature extraction ability of the model on the hard samples. Moreover, integrated gradient is introduced to attribute the contribution of each parameter in the alternate models trained by both sampling data so as to filter the trivial parameters for further dynamic refinement. By progressively decoupling the allocation of parameters, the SR model can learn a more compact representation. Extensive experiments on publicly available datasets demonstrate that our proposal can effectively boost the performance of baseline methods from the data re-sampling view.

Deep learning has made unprecedented progress in image restoration (IR), where residual block (RB) is popularly used and has a significant effect on promising performance. However, the massive stacked RBs bring about burdensome memory and computation cost. To tackle this issue, we aim to design an economical structure for adaptively connecting pair-wise RBs, thereby enhancing the model representation. Inspired by the topological structure of lattice filter in signal processing theory, we elaborately propose the lattice block (LB), where couple butterfly-style topological structures are utilized to bridge pair-wise RBs. Specifically, each candidate structure of LB relies on the combination coefficients learned through adaptive channel reweighting. As a basic mapping block, LB can be plugged into various IR models, such as image super-resolution, image denoising, image deraining, etc. It can avail the construction of lightweight IR models accompanying half parameter amount reduced, while keeping the considerable reconstruction accuracy compared with RBs. Moreover, a novel contrastive loss is exploited as a regularization constraint, which can further enhance the model representation without increasing the inference expenses. Experiments on several IR tasks illustrate that our method can achieve more favorable performance than other state-of-the-art models with lower storage and computation.

Hidden features in the neural networks usually fail to learn informative representation for 3D segmentation as supervisions are only given on output prediction, while this can be solved by omni-scale supervision on intermediate layers. In this paper, we bring the first omni-scale supervision method to 3D segmentation via the proposed gradual Receptive Field Component Reasoning (RFCR), where target Receptive Field Component Codes (RFCCs) is designed to record categories within receptive fields for hidden units in the encoder. Then, target RFCCs will supervise the decoder to gradually infer the RFCCs in a coarse-to-fine categories reasoning manner, and finally obtain the semantic labels. To purchase more supervisions, we also propose an RFCR-NL model with complementary negative codes (i.e., Negative RFCCs, NRFCCs) with negative learning. Because many hidden features are inactive with tiny magnitudes and make minor contributions to RFCC prediction, we propose Feature Densification with a centrifugal potential to obtain more unambiguous features, and it is in effect equivalent to entropy regularization over features. More active features can unleash the potential of omni-supervision method. We embed our method into three prevailing backbones, which are significantly improved in all three datasets on both fully and weakly supervised segmentation tasks and achieve competitive performances.

Low-rank self-representation based subspace learning has confirmed its great effectiveness in a broad range of applications. Nevertheless, existing studies mainly focus on exploring the global linear subspace structure, and cannot commendably handle the case where the samples approximately (i.e., the samples contain data errors) lie in several more general affine subspaces. To overcome this drawback, in this paper, we innovatively propose to introduce affine and nonnegative constraints into low-rank self-representation learning. While simple enough, we provide their underlying theoretical insight from a geometric perspective. The union of two constraints geometrically restricts each sample to be expressed as a convex combination of other samples in the same subspace. In this way, when exploring the global affine subspace structure, we can also consider the specific local distribution of data in each subspace. To comprehensively demonstrate the benefits of introducing two constraints, we instantiate three low-rank self-representation methods ranging from single-view low-rank matrix learning to multi-view low-rank tensor learning. We carefully design the solution algorithms to efficiently optimize the proposed three approaches. Extensive experiments are conducted on three typical tasks, including single-view subspace clustering, multi-view subspace clustering, and multi-view semi-supervised classification. The notably superior experimental results powerfully verify the effectiveness of our proposals.

Night-Time Scene Parsing (NTSP) is essential to many vision applications, especially for autonomous driving. Most of the existing methods are proposed for day-time scene parsing. They rely on modeling pixel intensity-based spatial contextual cues under even illumination. Hence, these methods do not perform well in night-time scenes as such spatial contextual cues are buried in the over-/under-exposed regions in night-time scenes. In this paper, we first conduct an image frequency-based statistical experiment to interpret the day-time and night-time scene discrepancies. We find that image frequency distributions differ significantly between day-time and night-time scenes, and understanding such frequency distributions is critical to NTSP problem. Based on this, we propose to exploit the image frequency distributions for night-time scene parsing. First, we propose a Learnable Frequency Encoder (LFE) to model the relationship between different frequency coefficients to measure all frequency components dynamically. Second, we propose a Spatial Frequency Fusion module (SFF) that fuses both spatial and frequency information to guide the extraction of spatial context features. Extensive experiments show that our method performs favorably against the state-of-the-art methods on the NightCity, NightCity+ and BDD100K-night datasets. In addition, we demonstrate that our method can be applied to existing day-time scene parsing methods and boost their performance on night-time scenes. The code is available at https://github.com/wangsen99/FDLNet.

Video super-resolution (VSR) is a fundamental and challenging task in computer vision. Many of the existing VSR works focus on how to effectively align neighboring frames to better incorporate temporal information, while little work is devoted to the important subsequent step of inter-frame information fusion, and the existing methods on frame fusion have shortcomings such as not being able to make full use of spatio-temporal information. In this work, we propose a Frame-by-frame Feedback Fusion Network (FFFN) for VSR tasks. By applying the feedback learning mechanism commonly existing in the human cognitive system to the frame fusion stage, FFFN can refine low-level representation of the fused frames with high-level information in a coarse-to-fine manner. Specifically, after the neighboring frames are aligned, we first rearrange them from near to far according to the distance from the reference frame in the temporal space, and then feed them one-by-one into a proposed recurrent structure called Feedback Fusion Module (FFM), which is then able to iteratively generate high-level representation of the fused frames with several Feature Refinement Groups (FRGs) and feedback connections. Finally, we design a Dual-path Residual Reconstruction Module (DRRM) to reconstruct the final high-resolution image. The proposed FFFN comes with a strong frame fusion and reconstruction ability, and extensive experiments on several benchmark data sets show that it achieves favorable performance against state-of-the-art methods.

Video anomaly detection aims to automatically identify unusual objects or behaviours by learning from normal videos. Previous methods tend to use simplistic reconstruction or prediction constraints, which leads to the insufficiency of learned representations for normal data. As such, we propose a novel bi-directional architecture with three consistency constraints to comprehensively regularize the prediction task from pixel-wise, cross-modal, and temporal-sequence levels. First, predictive consistency is proposed to consider the symmetry property of motion and appearance in forwards and backwards time, which ensures the highly realistic appearance and motion predictions at the pixel-wise level. Second, association consistency considers the relevance between different modalities and uses one modality to regularize the prediction of another one. Finally, temporal consistency utilizes the relationship of the video sequence and ensures that the predictive network generates temporally consistent frames. During inference, the pattern of abnormal frames is unpredictable and will therefore cause higher prediction errors. Experiments show that our method outperforms advanced anomaly detectors and achieves state-of-the-art results on UCSD Ped2, CUHK Avenue, and ShanghaiTech datasets.

Learning with limited labeled data is a long-standing problem. Among various solutions, episodic training progres-sively classifies a series of few-shot tasks and thereby is as-sumed to be beneficial for improving the model’s generalization ability. However, recent studies show that it is eveninferior to the baseline model when facing domain shift between base and novel classes. To tackle this problem, we pro-pose a domain-independent task-level self-supervised (TL-SS) method for cross-domain few-shot learning.TL-SS strategy promotes the general idea of label-based instance-levelsupervision to task-level self-supervision by augmenting mul-tiple views of tasks. Two regularizations on task consistencyand correlation metric are introduced to remarkably stabi-lize the training process and endow the generalization ability into the prediction model. We also propose a high-order associated encoder (HAE) being adaptive to various tasks.By utilizing 3D convolution module, HAE is able to generate proper parameters and enables the encoder to flexibly toany unseen tasks. Two modules complement each other andshow great promotion against state-of-the-art methods experimentally. Finally, we design a generalized task-agnostic test,where our intriguing findings highlight the need to re-think the generalization ability of existing few-shot approaches.

Generalized zero-shot learning (GZSL) requires a classifier trained on seen classes that can recognize objects from both seen and unseen classes. Due to the absence of unseen training samples, the classifier tends to bias towards seen classes. To mitigate this problem, feature generation based models are proposed to synthesize visual features for unseen classes. However, these features are generated in the visual feature space which lacks of discriminative ability. Therefore, some methods turn to find a better embedding space for the classifier training. They emphasize the inter-class relationships of seen classes, leading the embedding space overfitted to seen classes and unfriendly to unseen classes. Instead, in this paper, we propose an Intra-Class Compactness Enhancement method (ICCE) for GZSL. Our ICCE promotes intra-class compactness with inter-class separability on both seen and unseen classes in the embedding space and visual feature space. By promoting the intra-class relationships but the inter-class structures, we can distinguish different classes with better generalization. Specifically, we propose a Self-Distillation Embedding (SDE) module and a Semantic-Visual Contrastive Generation (SVCG) module. The former promotes intra-class compactness in the embedding space, while the latter accomplishes it in the visual feature space. The experiments demonstrate that our ICCE outperforms the state-of-the-art methods on four datasets and achieves competitive results on the remaining dataset.

To address the huge labeling cost in large-scale point cloud semantic segmentation, we propose a novel hybrid contrastive regularization (HybridCR) framework in weakly-supervised setting, which obtains competitive performance compared to its fully-supervised counterpart. Specifically, HybridCR is the first framework to leverage both point consistency and employ contrastive regularization with pseudo labeling in an end-to-end manner. Fundamentally, HybridCR explicitly and effectively considers the semantic similarity between local neighboring points and global characteristics of 3D classes. We further design a dynamic point cloud augmentor to generate diversity and robust sample views, whose transformation parameter is jointly optimized with model training. Through extensive experiments, HybridCR achieves significant performance improvement against the SOTA methods on both indoor and outdoor datasets, e.g., S3DIS, ScanNet-V2, Semantic3D, and SemanticKITTI.

Point cloud completion aims at completing geometric and topological shapes from a partial observation. However, some topology of the original shape is missing, existing methods directly predict the location of complete points, without predicting structured and topological information of the complete shape, which leads to inferior performance. To better tackle the missing topology part, we propose LAKe-Net, a novel topology-aware point cloud completion model by localizing aligned keypoints, with a novel Keypoints-Skeleton-Shape prediction manner. Specifically, our method completes missing topology using three steps: 1) Aligned Keypoint Localization. An asymmetric keypoint locator, including an unsupervised multi-scale keypoint detector and a complete keypoint generator, is proposed for localizing aligned keypoints from complete and partial point clouds. We theoretically prove that the detector can capture aligned keypoints for objects within a sub-category. 2) Surface-skeleton Generation. A new type of skeleton, named Surface-skeleton, is generated from keypoints based on geometric priors to fully represent the topological information captured from keypoints and better recover the local details. 3) Shape Refinement. We design a refinement subnet where multi-scale surface-skeletons are fed into each recursive skeleton-assisted refinement module to assist the completion process. Experimental results show that our method achieves the state-of-the-art performance on point cloud completion.

Cross-modality person re-identification is a newly rising and challenging problem, as there is a significant gap between the visible and infrared images. Though recent methods rapidly narrow the gap, the intra-modality variance is often ignored before inter-modality alignment. In this paper, we study this problem in the knowledge distillation perspective and design a self-mimic mutual-distillation method to reduce the discrepancy of each person from intra-modality feature alignment to cross-modality feature alignment. For intra-modality feature alignment, the self-mimic mechanism is implemented to simultaneously learn globally viewed, stable, and distinguish prototypes for each ID and minimize the intra-modality discrepancy. For inter-modality feature alignment, the mutual distillation is conducted to minimize the cross-modality distribution discrepancy of each person. Extensive experimental results on SYSU-MM01 and RegDB demonstrate that the proposed method achieves the best performance, outperforming state-of-the-art methods by a large margin without adding extra network parameters to the baseline. Especially, on the SYSU-MM01 dataset, our method achieves 64.8% Rank-1 and 60.2% mAP with significant gains over the latest related method.

Transformers have shown remarkable progress on computer vision tasks in the past year. Compared to their CNN counterparts, transformers usually need the help of distillation to achieve comparable results on middle or small sized datasets. Meanwhile, recent researches discover that when transformers are trained with supervised and self-supervised manner respectively, the captured patterns are quite different both qualitatively and quantitatively. These findings motivate us to introduce an self-supervised teaching assistant (SSTA) besides the commonly used supervised teacher to improve the performance of transformers. Specifically, we propose a head-level knowledge distillation method that selects the most important head of the supervised teacher and self-supervised teaching assistant, and let the student mimic the attention distribution of these two heads, so as to make the student focus on the relationship between tokens deemed by the teacher and the teacher assistant. Extensive experiments verify the effectiveness of SSTA and demonstrate that the proposed SSTA is a good compensation to the supervised teacher. Meanwhile, some analytical experiments towards multiple perspectives (e.g. prediction, shape bias, robustness, and transferability to downstream tasks) with supervised teachers, self-supervised teaching assistants and students are inductive and may inspire future researches.

Few-shot learning problem focuses on recognizing unseen classes given a few labeled images. In recent effort, more attention is paid to fine-grained feature embedding, ignoring the relationship among different distance metrics. In this paper, for the first time, we investigate the contributions of different distance metrics, and propose an adaptive fusion scheme, bringing significant improvements in few-shot classification. We start from a naive baseline of confidence summation and demonstrate the necessity of exploiting the complementary property of different distance metrics. By finding the competition problem among them, built upon the baseline, we propose an Adaptive Metrics Module (AMM) to decouple metrics fusion into metric-prediction fusion and metric-losses fusion. The former encourages mutual complementary, while the latter alleviates metric competition via multi-task collaborative learning. Based on AMM, we design a few-shot classification framework AMTNet, including the AMM and the Global Adaptive Loss (GAL), to jointly optimize the few-shot task and auxiliary self-supervised task, making the embedding features more robust. In the experiment, the proposed AMM achieves 2% higher performance than the naive metrics fusion module, and our AMTNet outperforms the state-of-the-arts on multiple benchmark datasets.

With the emergence of multi-modal datasets where LiDAR and camera are synchronized and calibrated, cross-modal Unsupervised Domain Adaptation (UDA) has attracted increasing attention because it reduces the laborious annotation of target domain samples. To alleviate the distribution gap between source and target domains, existing methods conduct feature alignment by using adversarial learning. However, it is well-known to be highly sensitive to hyperparameters and difficult to train. In this paper, we propose a novel model (Dual-Cross) that integrates Cross-Domain Knowledge Distillation (CDKD) and Cross-Modal Knowledge Distillation (CMKD) to mitigate domain shift. Specifically, we design the multi-modal style transfer to convert source image and point cloud to target style. With these synthetic samples as input, we introduce a target-aware teacher network to learn knowledge of the target domain. Then we present dual-cross knowledge distillation when the student is learning on source domain. CDKD constrains teacher and student predictions under same modality to be consistent. It can transfer target-aware knowledge from the teacher to the student, making the student more adaptive to the target domain. CMKD generates hybrid-modal prediction from the teacher predictions and constrains it to be consistent with both 2D and 3D student predictions. It promotes the information interaction between two modalities to make them complement each other. From the evaluation results on various domain adaptation settings, Dual-Cross significantly outperforms both uni-modal and cross-modal state-of-the-art methods.

Though single image super-resolution (SR) has witnessed incredible progress, the increasing model complexity impairs its applications in memory-limited devices. To solve this problem, prior arts have aimed to reduce the number of model parameters and sparsity has been exploited, which usually enforces the group sparsity constraint on the filter level and thus is not arbitrarily adjustable for satisfying the customized memory requirements. In this paper, we propose an individual kernel sparsity (IKS) method for memory-efficient and sparsity-adjustable image SR to aid deep network deployment in memory-limited devices. IKS performs model sparsity in the weight level that implicitly allocates the user-defined target sparsity to each individual kernel. To induce the kernel sparsity, a soft thresholding operation is used as a gating constraint for filtering the trivial weights. To achieve adjustable sparsity, a dynamic threshold learning algorithm is proposed, in which the threshold is updated by associated training with the network weight and is adaptively decayed with the guidance of the desired sparsity. This work essentially provides a dynamic parameter reassignment scheme with a given resource budget for an off-the-shelf SR model. Extensive experimental results demonstrate that IKS imparts considerable sparsity with negligible effect on SR quality. The code is available at: https://github.com/RaccoonDML/IKS.

Visible-Infrared Person Re-Identification (VI-ReID) has become an emerging task for night-time surveillance systems. In order to reduce the cross-modality discrepancy, previous works either align the features via metric learning or generate synthesized cross-modality images by Generative Adversary Network. However, feature-level alignment ignores the heterogeneous data itself while generative framework suffers from the low generation quality, limiting their applications. In this paper, we propose a dense contrastive learning framework (DCLNet), which performs pixel-to-pixel dense alignment acting on the intermediate representations, rather than the final deep feature. It is a new loss function that brings views of positive pixels with same semantic information closer in shallow representation space, whilst pushing views of negative pixels apart. It naturally provides additional dense supervision and captures fine-grained pixel correspondence, reducing the modality gap from a new perspective. To implement it, a Part Aware Parsing (PAP) module and a Semantic Rectification Module (SRM) are introduced to learn and refine a semantic-guided mask, allowing us to efficiently find positive pairs only requiring instance-level supervision. Extensive experiments on the public SYSU-MM01 and RegDB datasets demonstrate the superiority of our pipeline over state-of-the-arts. Code is available at https://github.com/sunhz0117/DCLNet.

Recently, transformer purely based on attention mechanism has been applied to a wide range of tasks and achieved impressive performance. Though extensive efforts have been made, there are still drawbacks to the transformer architecture which hinder its further applications: (i) the quadratic complexity brought by attention mechanism; (ii) barely incorporated inductive bias.

Multi-label image classification, which can be categorized into label-dependency and region-based methods, is a challenging problem due to the complex underlying object layouts. Although region-based methods are less likely to encounter issues with model generalizability than label-dependency methods, they often generate hundreds of meaningless or noisy proposals with non-discriminative information, and the contextual dependency among the localized regions is often ignored or over-simplified. This paper builds a unified framework to perform effective noisy-proposal suppression and to interact between global and local features for robust feature learning. Specifically, we propose category-aware weak supervision to concentrate on non-existent categories so as to provide deterministic information for local feature learning, restricting the local branch to focus on more high-quality regions of interest. Moreover, we develop a cross-granularity attention module to explore the complementary information between global and local features, which can build the high-order feature correlation containing not only global-to-local, but also local-to-local relations. Both advantages guarantee a boost in the performance of the whole network. Extensive experiments on two large-scale datasets (MS-COCO and VOC 2007) demonstrate that our framework achieves superior performance over state-of-the-art methods.

2D-3D unsupervised domain adaptation (UDA) tackles the lack of annotations in a new domain by capitalizing the relationship between 2D and 3D data. Existing methods achieve considerable improvements by performing cross-modality alignment in a modality-agnostic way, failing to exploit modality-specific characteristic for modeling complementarity. In this paper, we present self-supervised exclusive learning for cross-modal semantic segmentation under the UDA scenario, which avoids the prohibitive annotation. Specifically, two self-supervised tasks are designed, named "plane-to-spatial'' and "discrete-to-textured''. The former helps the 2D network branch improve the perception of spatial metrics, and the latter supplements structured texture information for the 3D network branch. In this way, modality-specific exclusive information can be effectively learned, and the complementarity of multi-modality is strengthened, resulting in a robust network to different domains. With the help of the self-supervised tasks supervision, we introduce a mixed domain to enhance the perception of the target domain by mixing the patches of the source and target domain samples. Besides, we propose a domain-category adversarial learning with category-wise discriminators by constructing the category prototypes for learning domain-invariant features. We evaluate our method on various multi-modality domain adaptation settings, where our results significantly outperform both uni-modality and multi-modality state-of-the-art competitors.

Cross-modality person re-identification (Re-ID) is more challenging than traditional visible Re-ID due to the huge cross-modality gap from heterogeneous images. To alleviate this problem, existing methods often utilize a dual path learning framework equipped with metric loss to learn discriminative features. Despite effectiveness, the inevitable degeneration of intra-modality discrimination by taking cross-modality discrimination into consideration is unsolvable. Such degeneration substantially hinders the model’s capability of further improving feature representations. To mitigate this degeneration, we propose a Dual Mutual Learning (DML) method for cross-modality Re-ID which conducts mutual learning between the cross-modality and each of two single modalities. We design a triple-branch deep model containing the RGB and IR branches and the cross-modality branch. The cross-modality branch is designed to learn modality-invariant feature subspace for appearance similarity measurement. Both the RGB branch and IR branch provide attention supervision information to the cross-modality branch for attention feature alignment so as to enhance the intra-modality discrimination. Experimental results on two standard benchmarks demonstrate DML is superior to state-of-the-art methods.

Multiview subspace clustering (MSC) has attracted growing attention due to the extensive value in various applications, such as natural language processing, face recognition, and time-series analysis. In this article, we are devoted to address two crucial issues in MSC: 1) high computational cost and 2) cumbersome multistage clustering. Existing MSC approaches, including tensor singular value decomposition (t-SVD)-MSC that has achieved promising performance, generally utilize the dataset itself as the dictionary and regard representation learning and clustering process as two separate parts, thus leading to the high computational overhead and unsatisfactory clustering performance. To remedy these two issues, we propose a novel MSC model called joint skinny tensor learning and latent clustering (JSTC), which can learn high-order skinny tensor representations and corresponding latent clustering assignments simultaneously. Through such a joint optimization strategy, the multiview complementary information and latent clustering structure can be exploited thoroughly to improve the clustering performance. An alternating direction minimization algorithm, which owns low computational complexity and can be run in parallel when solving several key subproblems, is carefully designed to optimize the JSTC model. Such a nice property makes our JSTC an appealing solution for large-scale MSC problems. We conduct extensive experiments on ten popular datasets and compare our JSTC with 12 competitors. Five commonly used metrics, including four external measures (NMI, ACC, F-score, and RI) and one internal metric (SI), are adopted to evaluate the clustering quality. The experimental results with the Wilcoxon statistical test demonstrate the superiority of the proposed method in both clustering performance and operational efficiency.

Multi-view subspace clustering is an effective method to partition data into their corresponding categories. Nevertheless, existing multi-view subspace clustering approaches generally operate in a purely unsupervised manner, while ignoring the valuable weakly supervised information that can be readily obtained in many practical applications. In this paper, we consider the weakly supervised form of sample pair constraints, and devote to promoting the performance of multi-view subspace clustering with the aid of such prior knowledge. To achieve this goal, inspired by the intrinsic block diagonal structure of ideal low-rank representation (LRR), we propose a novel regularization to integrate must-link, cannot-link and normalization constraints into a unified formulation. The proposed regularization can be regarded as a general description for sample pairwise constraints, and thus provides a flexible framework for multi-view semi-supervised subspace clustering task. Furthermore, we devise a contrained tensor representation learning (CTRL) model that takes advantage of our proposed regularization to facilitate the learning of the desired representation tensor. An efficient optimization algorithm based on alternating direction minimization strategy is carefully designed to solve the proposed CTRL model. Extensive experiments on eight challenging real-world datasets are conducted, and the results validate the effectiveness of our designed pairwise constraints regularization, as well as the superiority of the proposed CTRL model.

Boundary information plays a significant role in 2D image segmentation, while usually being ignored in 3D point cloud segmentation where ambiguous features might be generated in feature extraction, leading to misclassification in the transition area between two objects. In this paper, firstly, we propose a Boundary Prediction Module (BPM) to predict boundary points. Based on the predicted boundary, a boundary-aware Geometric Encoding Module (GEM) is designed to encode geometric information and aggregate features with discrimination in a neighborhood, so that the local features belonging to different categories will not be polluted by each other. To provide extra geometric information for boundary-aware GEM, we also propose a light-weight Geometric Convolution Operation (GCO), making the extracted features more distinguishing. Built upon the boundary-aware GEM, we build our network and test it on benchmarks like ScanNet v2, S3DIS. Results show our methods can significantly improve the baseline and achieve state-of-the-art performance.

Existing methods for large-scale point cloud semantic segmentation require expensive, tedious and error-prone manual point-wise annotation. Intuitively, weakly supervised training is a direct solution to reduce the labeling costs. However, for weakly supervised large-scale point cloud semantic segmentation, too few annotations will inevitably lead to ineffective learning of network. We propose an effective weakly supervised method containing two components to solve the above problem. Firstly, we construct a pretext task, \textit{i.e.,} point cloud colorization, with a self-supervised training manner to transfer the learned prior knowledge from a large amount of unlabeled point cloud to a weakly supervised network. In this way, the representation capability of the weakly supervised network can be improved by knowledge from a heterogeneous task. Besides, to generative pseudo label for unlabeled data, a sparse label propagation mechanism is proposed with the help of generated class prototypes, which is used to measure the classification confidence of unlabeled point. Our method is evaluated on large-scale point cloud datasets with different scenarios including indoor and outdoor. The experimental results show the large gain against existing weakly supervised methods and comparable results to fully supervised methods.

Hidden features in neural network usually fail to learn informative representation for 3D segmentation as supervisions are only given on output prediction, while this can be solved by omni-scale supervision on intermediate layers. In this paper, we bring the first omni-scale supervision method to point cloud segmentation via the proposed gradual Receptive Field Component Reasoning (RFCR), where target Receptive Field Component Codes (RFCCs) are designed to record categories within receptive fields for hidden units in the encoder. Then, target RFCCs will supervise the decoder to gradually infer the RFCCs in a coarse-to-fine categories reasoning manner, and finally obtain the semantic labels. Because many hidden features are inactive with tiny magnitude and make minor contributions to RFCC prediction, we propose a Feature Densification with a centrifugal potential to obtain more unambiguous features, and it is in effect equivalent to entropy regularization over features. More active features can further unleash the potential of our omni-supervision method. We embed our method into four prevailing backbones and test on three challenging benchmarks. Our method can significantly improve the backbones in all three datasets. Specifically, our method brings new state-of-the-art performances for S3DIS as well as Semantic3D and ranks the 1st in the ScanNet benchmark among all the point-based methods. Code is publicly available at https://github.com/azuki-miho/RFCR.

The Information Bottleneck (IB) provides an information theoretic principle for representation learning, by retaining all information relevant for predicting label while minimizing the redundancy. Though IB principle has been applied to a wide range of applications, its optimization remains a challenging problem which heavily relies on the accurate estimation of mutual information. In this paper, we present a new strategy, Variational Self-Distillation (VSD), which provides a scalable, flexible and analytic solution to essentially fitting the mutual information but without explicitly estimating it. Under rigorously theoretical guarantee, VSD enables the IB to grasp the intrinsic correlation between representation and label for supervised training. Further-more, by extending VSD to multi-view learning, we introduce two other strategies, Variational Cross-Distillation (VCD) and Variational Mutual-Learning (VML), which significantly improve the robustness of representation to view-changes by eliminating view-specific and task-irrelevant in-formation. To verify our theoretically grounded strategies, we apply our approaches to cross-modal person Re-ID, and conduct extensive experiments, where the superior performance against state-of-the-art methods are demonstrated. Our intriguing findings highlight the need to rethink the way to estimate mutual information.

Single image dehazing is a challenging ill-posed problem due to the severe information degeneration. However, existing deep learning based dehazing methods only adopt clear images as positive samples to guide the training of dehazing network while negative information is unexploited. Moreover, most of them focus on strengthening the dehazing network with an increase of depth and width, leading to a significant requirement of computation and memory. In this paper, we propose a novel contrastive regularization (CR) built upon contrastive learning to exploit both the information of hazy images and clear images as negative and positive samples, respectively. CR ensures that the restored image is pulled to closer to the clear image and pushed to far away from the hazy image in the representation space.Furthermore, considering trade-off between performance and memory storage, we develop a compact dehazing network based on autoencoder-like (AE) framework. It involves an adaptive mixup operation and a dynamic feature enhancement module, which can benefit from preserving information flow adaptively and expanding the receptive field to improve the network’s transformation capability, respectively. We term our dehazing network with autoencoder and contrastive regularization as AECR-Net. The extensive experiments on synthetic and real-world datasets demonstrate that our AECR-Net surpass the state-of-the-art approaches. The code is released in https://github.com/GlassyWu/AECR-Net.

Image-text matching is an important multi-modal task with massive applications. It tries to match the image and the text with similar semantic information. Existing approaches do not explicitly transform the different modalities into a common space. Meanwhile, the attention mechanism which is widely used in image-text matching models does not have supervision. We propose a novel attention scheme which projects the image and text embedding into a common space and optimises the attention weights directly towards the evaluation metrics. The proposed attention scheme can be considered as a kind of supervised attention and requiring no additional annotations. It is trained via a novel Discrete-continuous action space policy gradient algorithm, which is more effective in modelling complex action space than previous continuous action space policy gradient. We evaluate the proposed methods on two widely-used benchmark datasets: Flickr30k and MS-COCO, outperforming the previous approaches by a large margin.

Large-scale point cloud semantic segmentation has wide applications. Current popular researches mainly focus on fully supervised learning which demands expensive and tedious manual point-wise annotation. Weakly supervised learning is an alternative way to avoid this exhausting an-notation. However, for large-scale point clouds with few labeled points, the network is difficult to extract discriminative features for unlabeled points, as well as the regularization of topology between labeled and unlabeled points is usually ignored, resulting in incorrect segmentation results.To address this problem, we propose a perturbed self-distillation (PSD) framework. Specifically, inspired by self-supervised learning, we construct the perturbed branch and enforce the predictive consistency among the perturbed branch and original branch. In this way, the graph topology of the whole point cloud can be effectively established by the introduced auxiliary supervision, such that the in-formation propagation between the labeled and unlabeled points will be realized. Besides point-level supervision, we present a well-integrated context-aware module to explicitly regularize the affinity correlation of labeled points. Therefore, the graph topology of the point cloud can be further refined. The experimental results evaluated on three large-scale datasets show the large gain (3.0% on average) against recent weakly supervised methods and comparable results to some fully supervised methods.

Novelty detection is the task of identifying whether a new data point is considered to be an inlier or an outlier. Generative Adversarial Networks (GAN)-based methods suffer from mode dropping and unstable training issue, which poses the greatest threat to learn the target class distribution. To solve mode dropping issues, the nearest neighbor generator is designed to ensure that for every training image there exists a candidate generated image that is near to it at optimality. The generator considers the entire distribution of training data without mode dropping. To avoid the instability training issue, we consider capturing the distribution of the target class by non-adversarial strategy. In addition, to provide great image priors and fully diversity candidate samples for the generator, we also design a two-step mapping process. Finally, Experiments show that our model has clear superiority over cutting-edge novelty detectors and achieves state-of-the-art results on the datasets.

Few-shot learning, which attempts to predict unlabeled samples with only a few labeled samples, has drawn more and more attention. Though recent works have achieved promising progress, none of them have noticed to establish consistency among episodes, leading to the ambiguity in latent embedding space. In this paper, we propose a novel Cross-Modality Graph Neural Network (CMGNN) to uncover the associations among episodes for consistent global embedding. Since the semantic information induced from NLP is relatively fixed compared to visual information space, we leverage it to construct meta nodes for each category to guide the corresponding visual feature learning through GNN. Moreover, to ensure global embedding, a distance loss function is designed to force the visual nodes closer to their associated meta nodes to a greater extent. Extensive experiments and ablation studies on four benchmark datasets show its superiority over many SOTA comparison methods.

Few-shot learning (FSL), aiming to extract new knowledge from very small amount of labeled samples, has attracted noticeable attentions recently. However, most of existing methods often fail when facing huge domain shift between seen and unseen classes. We think this should be attributed to the episode strategy which ignore utilizing support samples to induct the test classes. So in this paper, for the first time, we propose a bilevel episode strategy (BL-ES) to train a inductive graph network (IGN) that learn to both comparison and induction. Specifically, first, outer episodes in BL-ES simulate the cross-domain few-shot tasks constantly, while inner episodes learn to drive IGN to induct the common features of test classes. Then, the propsoed IGN captures the correlation among all samples to update meta points of each category in induction module. Finally, we introduce a geometrical constraint term utilizing meta points into the training loss, to update the nodes and edges in feature space. This way improves the robustness of training process. Extensive experiments show that our framework outperforms the state-of-the-art FSL alternatives, and are more suitable for real-world applications.

Novelty detection is the process of determining whether a query example differs from the learned training distribution. Previous generative adversarial networks based methods and self-supervised approaches suffer from instability training, mode dropping, and low discriminative ability. We overcome such problems by introducing a novel decoder-encoder framework. Firstly, a generative network (decoder) learns the representation by mapping the initialized latent vector to an image. In particular, this vector is initialized by considering the entire distribution of training data to avoid the problem of mode-dropping. Secondly, a contrastive network (encoder) aims to ``learn to compare'' through mutual information estimation, which directly helps the generative network to obtain a more discriminative representation by using a negative data augmentation strategy. Extensive experiments show that our model has significant superiority over cutting-edge novelty detectors and achieves new state-of-the-art results on various novelty detection benchmarks, e.g. CIFAR10 and DCASE. Moreover, our model is more stable for training in a non-adversarial manner, compared to other adversarial based novelty detection methods.

Human beings have a great generalization ability to recognize a novel category by only seeing a few number of samples. This is because humans possess the ability to learn from the concepts that already exist in our minds. However, many existing few-shot approaches fail in addressing such a fundamental problem, {\it i.e.,} how to utilize the knowledge learned in the past to improve the prediction for the new task. In this paper, we present a novel purified memory mechanism that simulates the recognition process of human beings. This new memory updating scheme enables the model to purify the information from semantic labels and progressively learn consistent, stable, and expressive concepts when episodes are trained one by one. On its basis, a Graph Augmentation Module (GAM) is introduced to aggregate these concepts and knowledge learned from new tasks via a graph neural network, making the prediction more accurate. Generally, our approach is model-agnostic and computing efficient with negligible memory cost. Extensive experiments performed on several benchmarks demonstrate the proposed method can consistently outperform a vast number of state-of-the-art few-shot learning methods.

Face anti-spoofing approaches based on domain generalization (DG) have drawn growing attention due to their robustness for unseen scenarios. Previous methods treat each sample from multiple domains indiscriminately during the training process, and endeavor to extract a common feature space to improve the generalization. However, due to complex and biased data distribution, directly treating them equally will corrupt the generalization ability. To settle the issue, we propose a novel Dual Reweighting Domain Generalization (DRDG) framework which iteratively reweights the relative importance between samples to further improve the generalization. Concretely, Sample Reweighting Module is first proposed to identify samples with relatively large domain bias, and reduce their impact on the overall optimization. Afterwards, Feature Reweighting Module is introduced to focus on these samples and extract more domain-irrelevant features via a self-distilling mechanism. Combined with the domain discriminator, the iteration of the two modules promotes the extraction of generalized features. Extensive experiments and visualizations are presented to demonstrate the effectiveness and interpretability of our method against the state-of-the-art competitors.

Convolutional neural networks (CNNs) are highly successful for super-resolution (SR) but often require sophisticated architectures with heavy memory cost and computational overhead significantly restricts their practical deployments on resource-limited devices. In this paper, we proposed a novel contrastive self-distillation (CSD) framework to simultaneously compress and accelerate various off-the-shelf SR models. In particular, a channel-splitting super-resolution network can first be constructed from a target teacher network as a compact student network. Then, we propose a novel contrastive loss to improve the quality of SR images and PSNR/SSIM via explicit knowledge transfer. Extensive experiments demonstrate that the proposed CSD scheme effectively compresses and accelerates several standard SR models such as EDSR, RCAN and CARN. Code is available at https://github.com/Booooooooooo/CSD.

Despite recent progress on semantic segmentation, there still exist huge challenges in high or ultra-high resolution images semantic segmentation. Although the latest collaborative global-local semantic segmentation methods such as GLNet [4] and PPN [18] have achieved impressive results, they are inefficient and not fit for practical applications. Thus, in this paper, we propose a novel and efficient collaborative global-local framework on the basis of PPN named Faster-PPN for high or ultra-high resolution images semantic segmentation which makes a better trade-off between the efficient and effectiveness towards the real-time speed. Specially, we propose Dual Mutual Learning to improve the feature representation of global and local branches, which conducts knowledge distillation mutually between the global and local branches. Furthermore, we design the Pixel Proposal Fusion Module to conduct the fine-grained selection mechanism which further reduces the redundant pixels for fusion resulting in the improvement of inference speed. The experimental results on three challenging high or ultra-high resolution datasets DeepGlobe, ISIC and BACH demonstrate that Faster-PPN achieves the best performance on accuracy, inference speed and memory usage compared with state-of-the-art approaches. Especially, our method achieves real-time and near real-time speed with 36 FPS and 17.7 FPS on ISIC and DeepGlobe, respectively.

Time series clustering has attracted growing attention due to the abundant data accessible and extensive value in various applications. The unique characteristics of time series, including high-dimension, warping, and the integration of multiple elastic measures, pose challenges for the present clustering algorithms, most of which take into account only part of these difficulties. In this paper, we make an effort to simultaneously address all aforementioned issues in time series clustering under a unified multiple kernels clustering (MKC) framework. Specifically, we first implicitly map the raw time series space into multiple kernel spaces via elastic distance measure functions. In such high-dimensional spaces, we resort to the tensor constraint based self-representation subspace clustering approach, which involves the self-paced learning paradigm, to explore the essential low-dimensional structure of the data, as well as the high-order complementary information from different elastic kernels. The proposed approach can be extended to more challenging multivariate time series clustering scenario in a direct but elegant way. Extensive experiments on 85 univariate and 10 multivariate time series datasets demonstrate the significant superiority of the proposed approach beyond the baseline and several state-of-the-art MKC methods.

In this paper, a novelDeepMulti-viewJointClustering (DMJC) framework is proposed, where multiple deep embedded features, multi-view fusion mechanism, and clustering assignments can be learned simultaneously. Through the joint learning strategy, the clustering-friendly multi-view features and useful multi-view complementary information can be exploited effectively to improve the clustering performance. Under the proposed joint learning framework, we design two ingenious variants of deep multi-view joint clustering models, whose multi-view fusion is implemented by two kinds of simple yet effective schemes. The first model, called DMJC-S, performs multi-view fusion in an implicit way via a novel multi-view soft assignment distribution. The second model, termed DMJC-T, defines a novel multi-view auxiliary target distribution to conduct the multi-view fusion explicitly. Both DMJC-S and DMJC-T are optimized under a KL divergence objective. Experiments on eight challenging image datasets demonstrate the superiority of both DMJC-S and DMJC-T over single/multi-view baselines and the state-of-the-art multi-view clustering methods, which proves the effectiveness of the proposed DMJC framework. To the best of our knowledge, this is the first work to model the multi-view clustering in a deep joint framework, which will provide a meaningful thinking in unsupervised multi-view learning.

In this article, we propose a multiview self-representation model for nonlinear subspaces clustering. By assuming that the heterogeneous features lie within the union of multiple linear subspaces, the recent multiview subspace learning methods aim to capture the complementary and consensus from multiple views to boost the performance. However, in real-world applications, data feature usually resides in multiple nonlinear subspaces, leading to undesirable results. To this end, we propose a kernelized version of tensor-based multiview subspace clustering, which is referred to as Kt-SVD-MSC, to jointly learn self-representation coefficients in mapped high-dimensional spaces and multiple views correlation in unified tensor space. In view-specific feature space, a kernel-induced mapping is introduced for each view to ensure the separability of self-representation coefficients. In unified tensor space, a new kind of tensor low-rank regularizer is employed on the rotated self-representation coefficient tensor to preserve the global consistency across different views. We also derive an algorithm to efficiently solve the optimization problem with all the subproblems having closed-form solutions. Furthermore, by incorporating the nonnegative and sparsity constraints, the proposed method can be easily extended to a useful variant, meaning that several useful variants can be easily constructed in a similar way. Extensive experiments of the proposed method are tested on eight challenging data sets, in which a significant (even a breakthrough) advance over state-of-the-art multiview clustering is achieved.

Despite recent progress on the segmentation of high-resolution images, there exist an unsolved problem, i.e., the trade-off among the segmentation accuracy, memory resources and inference speed. So far, GLNet is introduced for high or ultra-resolution image segmentation, which has reduced the computational memory of the segmentation network. However, it ignores the importances of different cropped patches, and treats tiled patches equally for fusion with the whole image, resulting in high computational cost. To solve this problem, we introduce a patch proposal network (PPN) in this paper, which adaptively distinguishes the critical patches from the trivial ones to fuse with the whole image for refining segmentation. PPN is a classification network which alleviates network training burden and improves segmentation accuracy. We further embed PPN in a global-local segmentation network, instructing global branch and refinement branch to work collaboratively. We implement our method on four image datasets:DeepGlobe, ISIC, CRAG and Cityscapes, the first two are ultra-resolution image datasets and the last two are high-resolution image datasets. The experimental results show that our method achieves almost the best segmentation performance compared with the state-of-the-art segmentation methods and the inference speed is 12.9 fps on DeepGlobe and 10 fps on ISIC. Moreover, we embed PPN with the general semantic segmentation network and the experimental results on Cityscapes which contains more object classes demonstrate the generalization ability on general semantic segmentation.

State-of-the-art deep dehazing models are often difficult in training. Knowledge distillation paves a way to train a student network assisted by a teacher network. However, most knowledge distill methods are used for image classification and segmentation as well as object detection, and few investigate distilling image restoration and use different task for knowledge transfer. In this paper, we propose a knowledge-distill dehazing network which distills image dehazing with the heterogeneous task imitation. In our network, the teacher is an off-the-shelf auto-encoder network and is used for image reconstruction. The dehazing network is trained assisted by the teacher network with the process-oriented learning mechanism. The student network imitates the task of image reconstruction in the teacher network. Moreover, we design a spatial-weighted channel-attention residual block for the student image dehazing network to adaptively learn the content-aware channel level attention and pay more attention to the features for dense hazy regions reconstruction. To evaluate the effectiveness of the proposed method, we compare our method with several state-of-the-art methods on two synthetic and real-world datasets, as well as real hazy images.

Single-image super-resolution plays an important role in computer vision area. However, previous works using convolutional neural networks perform badly when reconstructing high frequency details, result in over-smooth and lacking of textural information in the output. At the same time, super-resolution computation always relays on convolutional neural networks with huge depth, which is super tricky to train and use. In this paper, we propose a novel network with better textural details in wavelet domain, which is composed of a feature extract layer, residual channel attention groups (RCAG) and a residual up-sampling layer based on inverse discrete wavelet transform. Channel attention and spatial attention layers are inserted into residual channel and spatial attention blocks (RCSAB), enhancing the learning of high frequency information with attention maps. Composed of a chain of RCSAB and a channel attention layer with short skip connection, RCAG is good at catching long-term high frequency information. Then the feature mapping component is composed of a chain of RCAG. Experiment shows that our method performs better than state-of-the-art methods on benchmark datasets in different scales.

With the development of adversarial attack in deep learning, it is critical for abnormal detector to not only discover the out-of-distribution samples but also provide defence against the adversarial attacker. Since few previous universal detector is known to work well on both tasks, we consider against both scenarios by constructing a robust and effective technique, where one sample could be regarded as the abnormal sample if it exhibits a higher image reconstruction error. Due to the training instability issues existed in previous generative adversarial networks (GANs) based methods, in this paper we propose a dual auxiliary autoencoder to make a tradeoff between the capability of generator and discriminator, leading to a more stable training process and high-quality image reconstruction. Moreover, to generate discriminative and robust latent representations, the mutual information estimator regarded as latent regularizer is adopted to extract the most unique information of target class. Overall, our generative dual adversarial network simultaneously optimizes the image reconstruction space and latent space to improve the performance. Experiments show that our model has the clear superiority over cutting edge semi-supervised abnormal detectors and achieves the state-of-the-art results on the datasets.

Despite recent great progress on semantic segmentation, there still exist huge challenges in medical ultra-resolution image segmentation. The methods based on multi-branch structure can make a good balance between computational burdens and segmentation accuracy. However, the fusion structure in these methods require to be designed elaborately to achieve desirable result, which leads to model redundancy. In this paper, we propose Meta Segmentation Network (MSN) to solve this challenging problem. With the help of meta-learning, the fusion module of MSN is quite simple but effective. MSN can fast generate the weights of fusion layers through a simple meta-learner, requiring only a few training samples and epochs to converge. In addition, to avoid learning all branches from scratch, we further introduce a particular weight sharing mechanism to realize a fast knowledge adaptation and share the weights among multiple branches, resulting in the performance improvement and significant parameters reduction. The experimental results on two challenging ultra-resolution medical datasets BACH and ISIC show that MSN achieves the best performance compared with the state-of-the-art approaches.

Besides local features, global information plays an essential role in semantic segmentation, while recent works usually fail to explicitly extract the meaningful global information and make full use of it. In this paper, we propose a SceneEncoder module to impose a scene-aware guidance to enhance the effect of global information. The module predicts a scene descriptor, which learns to represent the categories of objects existing in the scene and directly guides the point-level semantic segmentation through filtering out categories not belonging to this scene. Additionally, to alleviate segmentation noise in local region, we design a region similarity loss to propagate distinguishing features to their own neighboring points with the same label, leading to the enhancement of the distinguishing ability of point-wise features. We integrate our methods into several prevailing networks and conduct extensive experiments on benchmark datasets ScanNet and ShapeNet. Results show that our methods greatly improve the performance of baselines and achieve state-of-the-art performance.

Subspace clustering is an effective method that has been successfully applied to many applications. Here, we propose a novel subspace clustering model for multi-view data using a latent representation termed Latent Multi-View Subspace Clustering (LMSC). Unlike most existing single-view subspace clustering methods, which directly reconstruct data points using original features, our method explores underlying complementary information from multiple views and simultaneously seeks the underlying latent representation. Using the complementarity of multiple views, the latent representation depicts data more comprehensively than each individual view, accordingly making subspace representation more accurate and robust. We proposed two LMSC formulations: linear LMSC (lLMSC), based on linear correlations between latent representation and each view, and generalized LMSC (gLMSC), based on neural networks to handle general relationships. The proposed method can be efficiently optimized under the Augmented Lagrangian Multiplier with Alternating Direction Minimization (ALM-ADM) framework. Extensive experiments on diverse datasets demonstrate the effectiveness of the proposed method.

In this paper, we address the multiview nonlinear subspace representation problem. Traditional multiview subspace learning methods assume that the heterogeneous features of the data usually lie within the union of multiple linear subspaces. However, instead of linear subspaces, the data feature actually resides in multiple nonlinear subspaces in many real-world applications, resulting in unsatisfactory clustering performance. To overcome this, we propose a hyper-Laplacian regularized multilinear multiview self-representation model, which is referred to as HLR-M2VS, to jointly learn multiple views correlation and a local geometrical structure in a unified tensor space and view-specific self-representation feature spaces, respectively. In unified tensor space, a well-founded tensor low-rank regularization is adopted to impose on the self-representation coefficient tensor to ensure global consensus among different views. In view-specific feature space, hypergraph-induced hyper-Laplacian regularization is utilized to preserve the local geometrical structure embedded in a high-dimensional ambient space. An efficient algorithm is then derived to solve the optimization problem of the established model with theoretical convergence guarantee. Furthermore, the proposed model can be extended to semisupervised classification without introducing any additional parameters. An extensive experiment of our method is conducted on many challenging datasets, where a clear advance over state-of-the-art multiview clustering and multiview semisupervised classification approaches is achieved.

Existing successful person re-identification (Re-ID) models often employ the part-level representation to extract the fine-grained information, but commonly use the loss that is particularly designed for global features, ignoring the relationship between semantic parts. In this paper, we present a novel triplet loss that emphasizes the salient parts and also takes the consideration of alignment. This loss is based on the crossing-bing matching metric that also known as Wasserstein Distance. It measures how much effort it would take to move the embeddings of local features to align two distributions, such that it is able to find an optimal transport matrix to re-weight the distance of different local parts. The distributions in support of local parts is produced via a new attention mechanism, which is calculated by the inner product between high-level global feature and local features, representing the importance of different semantic parts w.r.t. identification. We show that the obtained optimal transport matrix can not only distinguish the relevant and misleading parts, and hence assign different weights to them, but also rectify the original distance according to the learned distributions, resulting in an elegant solution for the mis-alignment issue. Besides, the proposed method is easily implemented in most Re-ID learning system with end-to-end training style, and can obviously improve their performance. Extensive experiments and comparisons with recent Re-ID methods manifest the competitive performance of our method.

This paper presents a tensor multi-task model for person re-identification (Re-ID). Due to discrepancy among cameras, our approach regards Re-ID from multiple cameras as different but related classification tasks, each task corresponding to a specific camera. In each task, we distinguish the person identity as a one-vs-all linear classification problem, where one classifier is associated with a specific person. By constructing all classifiers into a task-specific projection matrix, the proposed method could utilize all the matrices to form a tensor structure, and jointly train all the tasks in a uniform tensor space. In this space, by assuming the features of the same person under different cameras are generated from a latent subspace, and different identities under the same perspective share similar patterns, the high-order correlations, not only across different tasks but also within a certain task, can be captured by utilizing a new type of low-rank tensor constraint. Therefore, the learned classifiers transform the original feature vector into the latent space, where feature distributions across cameras can be well-aligned. Moreover, this model can be incorporated into multiple visual features to boost the performance, and easily extended to the unsupervised setting. Extensive experiments and comparisons with recent Re-ID methods manifest the competitive performance of our method.

In this paper, we reduce the image dehazing problem to an image-to-image translation problem, and propose Enhanced Pix2pix Dehazing Network (EPDN), which generates a haze-free image without relying on the physical scattering model. EPDN is embedded by a generative adversarial network, which is followed by a well-designed enhancer. Inspired by visual perception global-first theory, the discriminator guides the generator to create a pseudo realistic image on a coarse scale, while the enhancer following the generator is required to produce a realistic dehazing image on the fine scale. The enhancer contains two enhancing blocks based on the receptive field model, which reinforces the dehazing effect in both color and details. The embedded GAN is jointly trained with the enhancer. Extensive experiment results on synthetic datasets and real-world datasets show that the proposed EPDN is superior to the state-of-the-art methods in terms of PSNR, SSIM, PI, and subjective visual effect.

Tremendous progress has been witnessed on single image super-resolution (SR), where existing deep SR models achieve impressive performance in objective criteria, e.g., PSNR and SSIM. However, most of the SR methods are limited in visual perception, for example, they look too smooth. Generative adversarial network (GAN) favors SR visual effects over most of the deep SR models but is poor in objective criteria. In order to trade off the objective and subjective SR performance, we design a joint-attention discriminator with which GAN improves the SR performance in PSNR and SSIM, as well as maintaining the visual effect compared with non-attention GAN based SR models. The joint-attention discriminator contains dense channel-wise attention and cross-layer attention blocks. The former is applied in the shallow layers of the discriminator for channel-wise weighting combination of feature maps. The latter is employed to select feature maps in some middle and deep layers for effective discrimination. Extensive experiments are conducted on six benchmark datasets and the experimental results show that our proposed discriminator combining with different generators can achieve more realistic visual performances.

The guided filter (GF) is a widely used smoothing tool in computer vision and image processing. However, to the best of our knowledge, few papers investigate the mathematical connection between this filter and the least-squares optimization. In this paper, we first interpret the guided filter as the cyclic coordinate descent (CCD) solver of a least-squares objective function. This discovery implies an extension approach to generalize the guided filter since we can change the least-squares objective function and define new filters as the first pass iteration of the CCD solver of modified objective functions. In addition, referring to the iterative minimizing procedure of the CCD, we can derive new rolling filtering schemes. So, we are reasonable to say that our discovery not only reveals an approach to design new GF-like filters adapting to specific requirements of applications but also offers thorough explanations for two rolling filtering schemes of the guided filter as well as the method to extend them. Experiments prove our new proposed filters and rolling filtering schemes could produce state-of-the-art results.

Multi-index fusion has demonstrated impressive performances in the retrieval task by integrating different visual representations in a unified framework. However, previous works mainly consider propagating similarities via a neighbor structure, ignoring the high-order information among different visual representations. In this paper, we propose a new multi-index fusion scheme for image retrieval. By formulating this procedure as a multilinear-based optimization problem, the complementary information hidden in different indexes can be explored more thoroughly. Specifically, we first build our multiple indexes from various visual representations. Then, a so-called index-specific functional matrix, which aims to propagate similarities, is introduced to update the original index. The functional matrices are then optimized in a unified tensor space to achieve a refinement, such that the relevant images can be pushed closer. The optimization problem can be efficiently solved by the augmented Lagrangian method with a theoretical convergence guarantee. Unlike the traditional multi-index fusion scheme, our approach embeds the multi-index subspace structure into the new indexes with sparse constraint and, thus, it has little additional memory consumption in the online query stage. Experimental evaluation on three benchmark datasets reveals that the proposed approach achieves state-of-the-art performance, that is, N-score 3.94 on UKBench, mAP 94.1% on Holiday, and 62.39% on Market-1501.

The high-dimensional convolution is widely used in various disciplines but has a serious performance problem due to its high computational complexity. Over the decades, people took a handmade approach to design fast algorithms for the Gaussian convolution. Recently, requirements for various non-Gaussian convolutions have emerged and are continuously getting higher. However, the handmade acceleration approach is no longer feasible for so many different convolutions since it is a time-consuming and painstaking job. Instead, we propose an Acceleration Network (AccNet) which turns the work of designing new fast algorithms to training the AccNet. This is done by: 1, interpreting splatting, blurring, slicing operations as convolutions; 2, turning these convolutions to $g$CP layers to build AccNet. After training, the activation function $g$ together with AccNet weights automatically define the new splatting, blurring and slicing operations. Experiments demonstrate AccNet is able to design acceleration algorithms for a ton of convolutions including Gaussian/non-Gaussian convolutions and produce state-of-the-art results.

Multitask learning has shown great potentiality for visual tracking under a particle filter framework. However, the recent multitask trackers, which exploit the similarity between all candidates by imposing group sparsity on the candidate representations, have a limitation in robustness due to the diverse sampling of candidates. To deal with this issue, we propose a discriminative reverse sparse tracker via weighted multitask learning. Our positive and negative templates are retained from the target observations and the background, respectively. Here, the templates are reversely represented via the candidates, and the representation of each positive template is viewed as a single task. Compared with existing multitask trackers, the proposed algorithm has the following advantages. First, we regularize the target representations with the ℓ2,1-norm to exploit the similarity shared by the positive templates, which is reasonable because of the target appearance consistency in the tracking process. Second, the valuable prior relationship between the candidates and the templates is introduced into the representation model by a weighted multitask learning scheme. Third, both target information and background information are integrated to generate discriminative scores for enhancing the proposed tracker. The experimental results on challenging sequences show that the proposed algorithm is effective and performs favorably against 12 state-of-the-art trackers.

An effective representation model, which aims to mine the most meaningful information in the data, plays an important role in visual tracking. Some recent particle-filter-based trackers achieve promising results by introducing the low-rank assumption into the representation model. However, their assumed low-rank structure of candidates limits the robustness when facing severe challenges such as abrupt motion. To avoid the above limitation, we propose a temporal restricted reverse-low-rank learning algorithm for visual tracking with the following advantages: 1) the reverse-low-rank model jointly represents target and background templates via candidates, which exploits the low-rank structure among consecutive target observations and enforces the temporal consistency of target in a global level; 2) the appearance consistency may be broken when target suffers from sudden changes. To overcome this issue, we propose a local constraint via 11,2 mixed-norm, which can not only ensures the local consistency of target appearance, but also tolerates the sudden changes between two adjacent frames; and 3) to alleviate the inference of unreasonable representation values due to outlier candidates, an adaptive weighted scheme is designed to improve the robustness of the tracker. By evaluating on 26 challenge video sequences, the experiments show the effectiveness and favorable performance of the proposed algorithm against 12 state-of-the-art visual trackers.

In this paper, we propose a new low-rank and sparse representation model for moving object detection. The model preserves the natural space-time structure of video sequences by representing them as three-way tensors. Then, it operates the low-rank background and sparse foreground decomposition in the tensor framework. On the one hand, we use the tensor nuclear norm to exploit the spatio-temporal redundancy of background based on the circulant algebra. On the other, we use the new designed saliently fused-sparse regularizer (SFS) to adaptively constrain the foreground with spatio-temporal smoothness. To refine the existing foreground smooth regularizers, the SFS incorporates the local spatio-temporal geometric structure information into the tensor total variation by using the 3D locally adaptive regression kernel (3D-LARK). What is more, the SFS further uses the 3D-LARK to compute the space-time motion saliency of foreground, which is combined with the l1norm and improves the robustness of foreground extraction. Finally, we solve the proposed model with globally optimal guarantee. Extensive experiments on challenging well-known data sets demonstrate that our method significantly outperforms the state-of-the-art approaches and works effectively on a wide range of complex scenarios.

We investigate the scalable image classification problem with a large number of categories. Hierarchical visual data structures are helpful for improving the efficiency and performance of large-scale multi-class classification. We propose a novel image classification method based on learning hierarchical inter-class structures. Specifically, we first design a fast algorithm to compute the similarity metric between categories, based on which a visual tree is constructed by hierarchical spectral clustering. Using the learned visual tree, a test sample label is efficiently predicted by searching for the best path over the entire tree. The proposed method is extensively evaluated on the ILSVRC2010 and Caltech 256 benchmark datasets. The experimental results show that our method obtains significantly better category hierarchies than other state-of-the-art visual tree-based methods and, therefore, much more accurate classification.

In this paper, we propose a new low-rank tensor model based on the circulant algebra, namely, twist tensor nuclear norm (t-TNN). The twist tensor denotes a three-way tensor representation to laterally store 2-D data slices in order. On one hand, t-TNN convexly relaxes the tensor multirank of the twist tensor in the Fourier domain, which allows an efficient computation using fast Fourier transform. On the other, t-TNN is equal to the nuclear norm of block circulant matricization of the twist tensor in the original domain, which extends the traditional matrix nuclear norm in a block circulant way. We test the t-TNN model on a video completion application that aims to fill missing values and the experiment results validate its effectiveness, especially when dealing with video recorded by a nonstationary panning camera. The block circulant matricization of the twist tensor can be transformed into a circulant block representation with nuclear norm invariance. This representation, after transformation, exploits the horizontal translation relationship between the frames in a video, and endows the t-TNN model with a more powerful ability to reconstruct panning videos than the existing state-of-the-art low-rank models.

Hyperspectral images (HSIs) are inevitably corrupted by mixture noise during their acquisition process, in which various kinds of noise, e.g., Gaussian noise, impulse noise, dead lines, and stripes, may exist concurrently. In this paper, mixture noise removal is well illustrated by the task of recovering the low-rank and sparse components of a given matrix, which is constructed by stacking vectorized HSI patches from all the bands at the same position. Instead of applying a traditional nuclear norm, a nonconvex low-rank regularizer, i.e., weighted Schatten p-norm (WSN), is introduced to not only give better approximation to the original low-rank assumption but also to consider the importance of different rank components. The resulted nonconvex low-rank matrix approximation (LRMA) model falls into the applicable scope of an augmented Lagrangian method, and its WSN minimization subproblem can be efficiently solved by generalized iterated shrinkage algorithm. Moreover, the proposed model is integrated into an iterative regularization schema to produce final results, leading to a completed HSI restoration framework. Extensive experimental testing on simulated and real data shows, both qualitatively and quantitatively, that the proposed method has achieved highly competent objective performance compared with several state-of-the-art HSI restoration methods.

Low rank matrix approximation (LRMA), which aims to recover the underlying low rank matrix from its degraded observation, has a wide range of applications in computer vision. The latest LRMA methods resort to using the nuclear norm minimization (NNM) as a convex relaxation of the nonconvex rank minimization. However, NNM tends to over-shrink the rank components and treats the different rank components equally, limiting its flexibility in practical applications. We propose a more flexible model, namely, the weighted Schatten p-norm minimization (WSNM), to generalize the NNM to the Schatten p-norm minimization with weights assigned to different singular values. The proposed WSNM not only gives better approximation to the original low-rank assumption, but also considers the importance of different rank components. We analyze the solution of WSNM and prove that, under certain weights permutation, WSNM can be equivalently transformed into independent non-convex lp-norm subproblems, whose global optimum can be efficiently solved by generalized iterated shrinkage algorithm. We apply WSNM to typical low-level vision problems, e.g., image denoising and background subtraction. Extensive experimental results show, both qualitatively and quantitatively, that the proposed WSNM can more effectively remove noise, and model the complex and dynamic scenes compared with state-of-the-art methods.

It remains a challenge to simultaneously remove geometric distortion and space-time-varying blur in frames captured through a turbulent atmospheric medium. To solve, or at least reduce these effects, we propose a new scheme to recover a latent image from observed frames by integrating a new hybrid total variation model and deformation-guided spatial-temporal kernel regression. The proposed scheme first constructs a high-quality reference image from the observed frames using low-rank decomposition. Then, to generate an improved registered sequence, the reference image is iteratively optimized using a variational model containing the combined regularization of local and non-local total variations. The proposed optimization algorithm efficiently solves this model with convergence guarantee. Next, to reduce blur variation, deformation-guided spatial-temporal kernel regression is carried out to fuse the registered sequence into one image by introducing the concept of the near-stationary patch. Applying a blind deconvolution algorithm to the fused image produces the final output. Extensive experimental testing shows, both qualitatively and quantitatively, that the proposed method can effectively alleviate distortion, and blur and recover details of the original scene compared to the state-of-the-art methods.

We propose a robust tracking algorithm based on local sparse coding with discriminative dictionary learning and new keypoint matching schema. This algorithm consists of two parts: the local sparse coding with online updated discriminative dictionary for tracking (SOD part), and the keypoint matching refinement for enhancing the tracking performance (KP part). In the SOD part, the local image patches of the target object and background are represented by their sparse codes using an over-complete discriminative dictionary. Such discriminative dictionary, which encodes the information of both the foreground and the background, may provide more discriminative power. Furthermore, in order to adapt the dictionary to the variation of the foreground and background during the tracking, an online learning method is employed to update the dictionary. The KP part utilizes refined keypoint matching schema to improve the performance of the SOD. With the help of sparse representation and online updated discriminative dictionary, the KP part are more robust than the traditional method to reject the incorrect matches and eliminate the outliers. The proposed method is embedded into a Bayesian inference framework for visual tracking. Experimental results on several challenging video sequences demonstrate the effectiveness and robustness of our approach.