Multivariate time series forecasting (MTSF) aims to learn temporal dynamics among variables to forecast future time series. Existing statistical and deep learning-based methods suffer from limited learnable parameters and small-scale training data. Recently, large language models (LLMs) combining time series with textual prompts have achieved promising performance in MTSF. However, we discovered that current LLM-based solutions fall short in learning disentangled embeddings. We introduce TimeCMA, an intuitive yet effective framework for MTSF via cross-modality alignment. Specifically, we present a dual-modality encoding with two branches: the time series encoding branch extracts disentangled yet weak time series embeddings, and the LLM-empowered encoding branch wraps the same time series with text as prompts to obtain entangled yet robust prompt embeddings. As a result, such a cross-modality alignment retrieves both disentangled and robust time series embeddings, ``the best of two worlds'', from the prompt embeddings based on time series and prompt modality similarities. As another key design, to reduce the computational costs from time series with their length textual prompts, we design an effective prompt to encourage the most essential temporal information to be encapsulated in the last token: only the last token is passed to downstream prediction. We further store the last token embeddings to accelerate inference speed. Extensive experiments on eight real datasets demonstrate that TimeCMA outperforms state-of-the-arts.

Sensors are commonly deployed to perceive the environment. However, due to the high cost, sensors are usually sparsely deployed. Kriging is the tailored task to infer the unobserved nodes (without sensors) using the observed nodes (with sensors). The essence of kriging task is transferability. Recently, several inductive spatio-temporal kriging methods have been proposed based on graph neural networks, being trained based on a graph built on top of observed nodes via pretext tasks such as masking nodes out and reconstructing them. However, the graph in training is inevitably much sparser than the graph in inference that includes all the observed and unobserved nodes. The learned pattern cannot be well generalized for inference, denoted as graph gap. To address this issue, we first present a novel Increment training strategy: instead of masking nodes (and reconstructing them), we add virtual nodes into the training graph so as to mitigate the graph gap issue naturally. Nevertheless, the empty-shell virtual nodes without labels could have bad-learned features and lack supervision signals. To solve these issues, we pair each virtual node with its most similar observed node and fuse their features together; to enhance the supervision signal, we construct reliable pseudo labels for virtual nodes. As a result, the learned pattern of virtual nodes could be safely transferred to real unobserved nodes for reliable kriging. We name our new Kriging model with Increment Training Strategy as KITS. Extensive experiments demonstrate that KITS consistently outperforms existing methods by large margins, e.g., the improvement over MAE score could be as high as 18.33%.

Adapting generative models to specific domains presents an effective solution for satisfying specialized requirements. However, adapting to some complex domains remains challenging, especially when these domains require substantial paired data to capture the targeted distributions. Since unpaired data from a single modality, such as vision or language, is more readily available, we utilize the bidirectional mappings between vision and language learned by the unified generative model to enable training on unpaired data for domain adaptation. Specifically, we propose DoraCycle, which integrates two multimodal cycles: text-to-image-to-text and image-to-text-to-image. The model is optimized through cross-entropy loss computed at the cycle endpoints, where both endpoints share the same modality. This facilitates self-evolution of the model without reliance on annotated text-image pairs. Experimental results demonstrate that for tasks independent of paired knowledge, such as stylization, DoraCycle can effectively adapt the unified model using only unpaired data. For tasks involving new paired knowledge, such as specific identities, a combination of a small set of paired image-text examples and larger-scale unpaired data is sufficient for effective domain-oriented adaptation. The code will be released at https://github.com/showlab/DoraCycle.

Individual Large Language Models (LLMs) have demonstrated significant capabilities across various domains, such as healthcare and law.Recent studies also show that coordinated multi-agent systems exhibit enhanced decisionmaking and reasoning abilities through collaboration.However, due to the vulnerabilities of individual LLMs and the difficulty of accessing all agents in a multi-agent system, a key question arises: If attackers only know one agent, could they still generate adversarial samples capable of misleading the collective decision?To explore this question, we formulate it as a game with incomplete information, where attackers know only one target agent and lack knowledge of the other agents in the system.With this formulation, we propose M-Spoiler, a framework that simulates agent interactions within a multi-agent system to generate adversarial samples.These samples are then used to manipulate the target agent in the target system, misleading the system's collaborative decisionmaking process.More specifically, M-Spoiler introduces a stubborn agent that actively aids in optimizing adversarial samples by simulating potential stubborn responses from agents in the target system.This enhances the effectiveness of the generated adversarial samples in misleading the system.Through extensive experiments across various tasks, our findings confirm the risks posed by the knowledge of an individual agent in multi-agent systems and demonstrate the effectiveness of our framework.We also explore several defense mechanisms, showing that our proposed attack framework remains more potent than baselines, underscoring the need for further research into defensive strategies.Our source code is available at here.

Graph neural networks (GNNs) face scalability challenges. While recent approaches have adopted pooling strategies inspired by convolutional neural networks (CNNs) to reduce graph size and improve efficiency, these methods often focus on local information and are optimized for single graph-level tasks. This limitation hinders their effectiveness in multi-task scenarios that require task-specific global information. We present DOTIN (Dropping Out Task-Irrelevant Nodes), an approach to graph size reductio. DOTIN utilizes K learnable virtual nodes to represent graph embeddings for K distinct graph-level tasks. By employing a transformer-based attention model, it adaptively removes up to 90% of low-attentiveness raw nodes without notable performance degradation. Our method achieves comparable accuracy to state-of-the-art techniques while offering substantial benefits in efficiency. Specifically, DOTIN accelerates Graph Attention Networks (GAT) by approximately 50% on graph-level tasks such as graph classification and graph edit distance (GED). Additionally, it reduces memory usage by about 60% on the D&D dataset. These results show DOTIN's potential to enhance the scalability and efficiency of deep GNNs across multiple graph-level tasks while maintaining high performance.

Recent advancements in multimodal foundation models have yielded significant progress in vision-language understanding. Initial attempts have also explored the potential of multimodal large language models (MLLMs) for visual content generation. However, existing works have insufficiently addressed the varying granularity demands of different image generation tasks within a unified MLLM paradigm - from the diversity required in text-to-image generation to the precise controllability needed in image manipulation. In this work, we propose PUMA, emPowering Unified MLLM with Multi-grAnular visual generation. PUMA unifies multi-granular visual features as both inputs and outputs of MLLMs, elegantly addressing the different granularity requirements of various image generation tasks within a unified MLLM framework. Following multimodal pretraining and task-specific instruction tuning, PUMA demonstrates proficiency in a wide range of multimodal tasks. This work represents a significant step towards a truly unified MLLM capable of adapting to the granularity demands of various visual tasks. The code and model will be released in https://github.com/rongyaofang/PUMA.

Multivariate time series forecasting (MTSF) endeavors to predict future observations given historical data, playing a crucial role in time series data management systems. With advancements in large language models (LLMs), recent studies employ textual prompt tuning to infuse the knowledge of LLMs into MTSF. However, the deployment of LLMs often suffers from low efficiency during the inference phase. To address this problem, we introduce TimeKD, an efficient MTSF framework that leverages the calibrated language models and privileged knowledge distillation. TimeKD aims to generate high-quality future representations from the proposed cross-modality teacher model and cultivate an effective student model. The cross-modality teacher model adopts calibrated language models (CLMs) with ground truth prompts, motivated by the paradigm of Learning Under Privileged Information (LUPI). In addition, we design a subtractive cross attention (SCA) mechanism to refine these representations. To cultivate an effective student model, we propose an innovative privileged knowledge distillation (PKD) mechanism including correlation and feature distillation. PKD enables the student to replicate the teacher's behavior while minimizing their output discrepancy. Extensive experiments on real data offer insight into the effectiveness, efficiency, and scalability of the proposed TimeKD.

Language-based object detection (LOD) aims to align visual objects with language expressions. A large amount of paired data is utilized to improve LOD model generalizations. During the training process, recent studies leverage vision-language models (VLMs) to automatically generate human-like expressions for visual objects, facilitating training data scaling up. In this process, we observe that VLM hallucinations bring inaccurate object descriptions (e.g., object name, color, and shape) to deteriorate VL alignment quality. To reduce VLM hallucinations, we propose an agentic workflow controlled by an LLM to re-align language to visual objects via adaptively adjusting image and text prompts. We name this workflow Real-LOD, which includes planning, tool use, and reflection steps. Given an image with detected objects and VLM raw language expressions, Real-LOD reasons its state automatically and arranges action based on our neural symbolic designs (i.e., planning). The action will adaptively adjust the image and text prompts and send them to VLMs for object re-description (i.e., tool use). Then, we use another LLM to analyze these refined expressions for feedback (i.e., reflection). These steps are conducted in a cyclic form to gradually improve language descriptions for re-aligning to visual objects. We construct a dataset that contains a tiny amount of 0.18M images with re-aligned language expression and train a prevalent LOD model to surpass existing LOD methods by around 50% on the standard benchmarks. Our Real-LOD workflow, with automatic VL refinement, reveals a potential to preserve data quality along with scaling up data quantity, which further improves LOD performance from a data-alignment perspective.

Few-Shot Segmentation (FSS) aims to learn class-agnostic segmentation on few classes to segment arbitrary classes, but at the risk of overfitting. To address this, some methods use the well-learned knowledge of foundation models (e.g., SAM) to simplify the learning process. Recently, SAM 2 has extended SAM by supporting video segmentation, whose class-agnostic matching ability is useful to FSS. A simple idea is to encode support foreground (FG) features as memory, with which query FG features are matched and fused. Unfortunately, the FG objects in different frames of SAM 2’s video data are always the same identity, while those in FSS are different identities, i.e., the matching step is incompatible. Therefore, we design Pseudo Prompt Generator to encode pseudo query memory, matching with query features in a compatible way. However, the memories can never be as accurate as the real ones, i.e., they are likely to contain incomplete query FG, and some unexpected query background (BG) features, leading to wrong segmentation. Hence, we further design Iterative Memory Refinement to fuse more query FG features into the memory, and devise a Support-Calibrated Memory Attention to suppress the unexpected query BG features in memory. Extensive experiments have been conducted on PASCAL-5$^i$ and COCO-20$^i$ to validate the effectiveness of our design, e.g., the 1-shot mIoU can be 4.2% better than the best baseline.

The proliferation of edge devices has generated an unprecedented volume of time series data across different domains, motivating a variety of well-customized methods. Recently, Large Language Models (LLMs) have emerged as a new paradigm for time series analytics by leveraging the shared sequential nature of textual data and time series. However, a fundamental cross-modality gap between time series and LLMs exists, as LLMs are pre-trained on textual corpora and are not inherently optimized for time series. Many recent proposals are designed to address this issue. In this survey, we provide an up-to-date overview of LLMs-based cross-modality modeling for time series analytics. We first introduce a taxonomy that classifies existing approaches into four groups based on the type of textual data employed for time series modeling. We then summarize key cross-modality strategies, e.g., alignment and fusion, and discuss their applications across a range of downstream tasks. Furthermore, we conduct experiments on multimodal datasets from different application domains to investigate effective combinations of textual data and cross-modality strategies for enhancing time series analytics. Finally, we suggest several promising directions for future research. This survey is designed for a range of professionals, researchers, and practitioners interested in LLM-based time series modeling.

Preference-based reinforcement learning (PbRL) bypasses complex reward engineering by learning from human feedback. However, due to the high cost of obtaining feedback, PbRL typically relies on a limited set of preference-labeled samples. This data scarcity introduces two key inefficiencies: (1) the reward model overfits to the limited feedback, leading to poor generalization to unseen samples, and (2) the agent exploits the learned reward model, exacerbating overestimation of action values in temporal difference (TD) learning. To address these issues, we propose STAR, an efficient PbRL method that integrates preference margin regularization and policy regularization. Preference margin regularization mitigates overfitting by introducing a bounded margin in reward optimization, preventing excessive bias toward specific feedback. Policy regularization bootstraps a conservative estimate $\widehat{Q}$ from well-supported state-action pairs in the replay memory, reducing overestimation during policy learning. Experimental results show that STAR improves feedback efficiency, achieving 34.8\% higher performance in online settings and 29.7\% in offline settings compared to state-of-the-art methods. Ablation studies confirm that STAR facilitates more robust reward and value function learning. The videos of this project are released at https://sites.google.com/view/pbrl-star.

Recently, person re-identification (ReID) has witnessed fast development due to its broad practical applications and proposed various settings, e.g., traditional ReID, clothes-changing ReID, and visible-infrared ReID. However, current studies primarily focus on single specific tasks, which limits model applicability in real-world scenarios. This paper aims to address this issue by introducing a novel instruct-ReID task that unifies 6 existing ReID tasks in one model and retrieves images based on provided visual or textual instructions. Instruct-ReID is the first exploration of a general ReID setting, where 6 existing ReID tasks can be viewed as special cases by assigning different instructions. To facilitate research in this new instruct-ReID task, we propose a large-scale OmniReID++ benchmark equipped with diverse data and comprehensive evaluation methods, e.g., task-specific and task-free evaluation settings. In the task-specific evaluation setting, gallery sets are categorized according to specific ReID tasks. We propose a novel baseline model, IRM, with an adaptive triplet loss to handle various retrieval tasks within a unified framework. For task-free evaluation setting, where target person images are retrieved from task-agnostic gallery sets, we further propose a new method called IRM++ with novel memory bank-assisted learning. Extensive evaluations of IRM and IRM++ on OmniReID++ benchmark demonstrate the superiority of our proposed methods, achieving state-of-the-art performance on 10 test sets.

Human-centric perception tasks, e.g., pedestrian detection, skeleton-based action recognition, and pose estimation, have wide industrial applications, such as metaverse and sports analysis. There is a recent surge to develop human-centric foundation models that can benefit a broad range of human-centric perception tasks. While many human-centric foundation models have achieved success, they did not explore 3D and vision-language tasks for human-centric and required task-specific finetuning. These limitations restrict their application to more downstream tasks and situations. To tackle these problems, we present Hulk, the first multimodal human-centric generalist model, capable of addressing 2D vision, 3D vision, skeleton-based, and vision-language tasks without task-specific finetuning. The key to achieving this is condensing various task-specific heads into two general heads, one for discrete representations, e.g., languages, and the other for continuous representations, e.g., location coordinates. The outputs of two heads can be further stacked into four distinct input and output modalities. This uniform representation enables Hulk to treat diverse human-centric tasks as modality translation, integrating knowledge across a wide range of tasks. Comprehensive evaluations of Hulk on 12 benchmarks covering 8 human-centric tasks demonstrate the superiority of our proposed method, achieving state-of-the-art performance in 11 benchmarks.

Visual relationships are crucial for visual perception and reasoning, and cover tasks like Scene Graph Generation, Human-Object Interaction, and object affordance. Despite significant efforts, this field still suffers from the following limitations: specialists for a specific task without considering similar ones, strict and complex task formulations with limited flexibility, and underexploited reasoning with language and knowledge. To solve these limitations, we seek to build a new framework, one model for all tasks, over Large Multimodal Models (LMMs). LMMs offer the potential of unifying tasks, flexible forms, and reasoning with language. However, they fail to handle visual relationship tasks well. We find the obstacles include the conflicts between different tasks and insufficient instance-level information. We solve these problems by reforming the data for LMMs, rather than architectures, considering their strong language-in language-out capability. We propose to disassemble tasks into simple and common sub-tasks, verbally estimate instance confidence, and augment instance diversity, all without additional modules. These strategies help us build a visual relationship generalist, RelationLMM, with a simple architecture. Exhaustive experiments demonstrate RelationLMM is strong, generalizable and flexible to different tasks, with one model and one suite of weight.

Many robotics and industry applications have a high demand for the capability to estimate the 6D pose of novel objects from the cluttered scene. However, existing classic pose estimation methods are object-specific, which can only handle the specific objects seen during training. When applied to a novel object, these methods necessitate a cumbersome onboarding process, which involves extensive dataset preparation and model retraining. The extensive duration and resource consumption of onboarding limit their practicality in real-world applications In this paper, we introduce ZeroPose, a novel zero-shot framework that performs pose estimation following a Discovery-Orientation-Registration (DOR) inference pipeline. This framework generalizes to novel objects without requiring model retraining. Given the CAD model of a novel object, ZeroPose enables in seconds onboarding time to extract visual and geometric embeddings from the CAD model as a prompt. With the prompting of the above embeddings, DOR can discover all related instances and estimate their 6D poses without additional human interaction or presupposing scene conditions. Compared with existing zero-shot methods solved by the render-and-compare paradigm, the DOR pipeline formulates the object pose estimation into a feature-matching problem, which avoids time-consuming online rendering and improves efficiency. Experimental results on the seven datasets show that ZeroPose as a zero-shot method achieves comparable performance with object-specific training methods and outperforms the state-of-the-art zero-shot method with 50x inference speed improvement.

Trajectory recovery from snapshots captured by a city-wide multi-camera network facilitates urban mobility sensing and road network optimization. State-of-the-art solutions for such vision-based schemes typically rely on predefined rules or unsupervised iterative feedback, but they struggle with multiple challenges, such as the lack of open-source datasets for training the entire pipeline and the vulnerability to noise in visual inputs. In response to the dilemma, this paper proposes VisionTraj, the first learning-based model that reconstructs vehicle trajectories from snapshots recorded by road network cameras. Along with this, we present two well-designed vision-trajectory datasets that provide extensive trajectory data and corresponding visual snapshots, enabling the extraction of supervised vision-trajectory interactions. After the data creation, based on the results from the off-the-shelf multi-modal vehicle clustering, we first re-formulate the trajectory recovery problem as a generative task and introduce the canonical Transformer as the autoregressive backbone. Next, to identify clustering noise (i.e., false positives) based on the snapshots’ spatiotemporal dependencies, a graph convolutional neural network-based soft-denoising module is built upon the fine- and coarse-grained clusters. Additionally, we leverage strong semantic information extracted from the tracklet to provide detailed insights into the vehicle’s entry and exit behaviors during trajectory recovery. The denoising and tracklet components can also serve as plug-and-play modules to enhance baselines. Experimental results on the two hand-crafted datasets show that the proposed VisionTraj achieves a maximum improvement of +11.5% against the sub-best model. Furthermore, we explore potential downstream applications, and our model continues to outperform its peers. The code and data are available herehttps://github.com/bonaldli/VisionTraj

Traffic prediction is a crucial component of data management systems, leveraging historical data to learn spatio-temporal dynamics for forecasting future traffic and enabling efficient decision-making and resource allocation. Despite efforts to develop increasingly complex architectures, existing traffic prediction models often struggle to generalize across diverse datasets and contexts, limiting their adaptability in real-world applications. In contrast to existing traffic prediction models, large language models (LLMs) progress mainly through parameter expansion and extensive pre-training while maintaining their fundamental structures. In this paper, we propose ST-LLM+, the graph enhanced spatio-temporal large language models for traffic prediction. Through incorporating a proximity-based adjacency matrix derived from the traffic network into the calibrated LLMs, ST-LLM+ captures complex spatio-temporal dependencies within the traffic network. The Partially Frozen Graph Attention (PFGA) module is designed to retain global dependencies learned during LLMs pre-training while modeling localized dependencies specific to the traffic domain. To reduce computational overhead, ST-LLM+ adopts the LoRA-augmented training strategy, allowing attention layers to be fine-tuned with fewer learnable parameters. Comprehensive experiments on real-world traffic datasets demonstrate that ST-LLM+ outperforms state-of-the-art models. In particular, ST-LLM+ also exhibits robust performance in both few-shot and zero-shot prediction scenarios. Additionally, our case study demonstrates that ST-LLM+ captures global and localized dependencies between stations, verifying its effectiveness for traffic prediction tasks.

Direct pose estimation networks aim to directly regress the 6D poses of target objects in the scene image using a neural network. These direct methods offer efficiency and an optimal optimization target, presenting significant potential for practical applications. However, due to the complex and implicit mappings between input features and target pose parameters, direct methods are challenging to train and prone to overfitting on mappings seen during training, resulting in limited effectiveness and generalization capability on unseen mappings. Existing methods focus primarily on improvements of the network architecture and training strategies, with less attention given to mappings. In this work, we propose a geometric constraints learning approach, which enables networks to explicitly capture and utilize the geometric mappings between inputs and optimization targets for pose estimation. Specifically, we introduce a residual pose transformation formula that preserves pose transformation constraints within both the 2D image plane and the 3D space while decoupling the absolute pose distribution, thereby addressing the pose distribution gap issue. We further design a Geo6D mechanism based on the formula, which enables the network to explicitly utilize geometric constraints for pose estimation by reconstructing the inputs and outputs. We select two different methods as our baseline and extensive experiments show that Geo6D enhances the performance and reduces the dependence on extensive training data, remaining effective even with only 10% of the typical data volume.

Kriging aims to estimate the attributes of unseen geo-locations from observations in the spatial vicinity or physical connections. Existing works assume that neighbors’ information offers the basis for estimating the unobserved target while ignoring non-neighbors. However, neighbors could also be quite different or even misleading, and the non-neighbors could still offer constructive information. To this end, we propose "Contrastive-Prototypical" self-supervised learning for Kriging (KCP): (1) The neighboring contrastive module coarsely pushes neighbors together and non-neighbors apart. (2) In parallel, the prototypical module identifies similar representations via exchanged prediction, such that it refines the misleading neighbors and recycles the useful non-neighbors from the neighboring contrast component. As a result, not all the neighbors and some of the non-neighbors will be used to infer the target. (3) To learn general and robust representations, we design an adaptive augmentation module that encourages data diversity. Theoretical bound is derived for the proposed augmentation. Extensive experiments on real-world datasets demonstrate the superior performance of KCP compared to its peers with 6% improvements and exceptional transferability and robustness.

Current diffusion-based video editing primarily focuses on structure-preserved editing by utilizing various dense correspondences to ensure temporal consistency and motion alignment. However, these approaches are often in-effective when the target edit involves a shape change. To embark on video editing with shape change, we explore customized video subject swapping in this work, where we aim to replace the main subject in a source video with a target subject having a distinct identity and potentially different shape. In contrast to previous methods that rely on dense correspondences, we introduce the Video Swap framework that exploits semantic point correspondences, inspired by our observation that only a small number of semantic points are necessary to align the subject's motion trajectory and modify its shape. We also introduce various user-point interactions (e.g., removing points and dragging points) to address various semantic point correspondence. Extensive experiments demonstrate state-of-the-art video subject swapping results across a variety of real-world videos.

Human intelligence can retrieve any person according to both visual and language descriptions. However, the current computer vision community studies specific person re-identification (ReID) tasks in different scenarios separately, which limits the applications in the real world. This paper strives to resolve this problem by proposing a new instruct-ReID task that requires the model to retrieve images according to the given image or language instructions. Our instruct-ReID is a more general ReID setting, where existing 6 ReID tasks can be viewed as special cases by designing different instructions. We propose a large-scale OmniReID benchmark and an adaptive triplet loss as a baseline method to facilitate research in this new setting. Experimental results show that the proposed multi-purpose ReID model, trained on our OmniReID benchmark without finetuning, can improve +0.5%, +0.6%, +7.7% mAP on Market1501, MSMT17, CUHK03 for traditional ReID, +6.4%, +7.1%, +11.2% mAP on PRCC, VC-Clothes, LTCC for clothes-changing ReID, +11.7% mAP on COCAS+ real2 for clothes template based clothes-changing ReID when using only RGB images, +24.9% mAP on COCAS+ real2 for our newly defined language-instructed ReID, +4.3% on LLCM for visible-infrared ReID, +2.6% on CUHK-PEDES for text-to-image ReID. The datasets, the model, and code are available at https://github.com/hwz-zju/Instruct-ReID.

Current self-supervised methods can primarily be categorized into contrastive learning and masked image modeling. Extensive studies have demonstrated that combining these two approaches can achieve state-of-the-art performance. However, these methods essentially reinforce the global consistency of contrastive learning without taking into account the conflicts between these two approaches, which hinders their generalizability to arbitrary scenarios. In this paper, we theoretically prove that MAE serves as a patch-level contrastive learning, where each patch within an image is considered as a distinct category. This presents a significant conflict with global-level contrastive learning, which treats all patches in an image as an identical category. To address this conflict, this work abandons the non-generalizable global-level constraints and proposes explicit patch-level contrastive learning as a solution. Specifically, this work employs the encoder of MAE to generate dual-branch features, which then perform patch-level learning through a decoder. In contrast to global-level data aug-mentation in contrastive learning, our approach leverages patch-level feature augmentation to mitigate interference from global-level learning. Consequently, our approach can learn heterogeneous representations from a single image while avoiding the conflicts encountered by previous methods. Massive experiments affirm the potential of our method for learning from arbitrary scenarios.

Despite recent progress in diffusion-based video editing, existing methods are limited to short-length videos due to the contradiction between long-range consistency and frame-wise editing. Prior attempts to address this challenge by introducing video-2D representations encounter significant difficulties with large motion- and view-change videos, especially in human-centric scenarios. To overcome this, we propose to introduce the dynamic Neural Radiance Fields (NeRF) as the innovative video representation, where the editing can be performed in the 3D spaces and propagated to the entire video via the deformation field. To provide consistent and controllable editing, we propose the image-based video-NeRF editing pipeline with a set of innovative designs, including multi-view multi-pose Score Distillation Sampling (SDS) from both the 2D personalized diffusion prior and 3D diffusion prior, reconstruction losses, text-guided local parts super-resolution, and style transfer. Extensive experiments demonstrate that our method dubbed as DynVideo-E, significantly outperforms SOTA approaches on two challenging datasets by a large margin of 50% ~ 95% for human preference. Code will be released at https://showlab.github.io/DynVideo-E/.

Large motion poses a critical challenge in Video Frame Interpolation (VFI) task. Existing methods are often constrained by limited receptive fields, resulting in sub-optimal performance when handling scenarios with large motion. In this paper, we introduce a new pipeline for VFI, which can effectively integrate global-level information to alleviate issues associated with large motion. Specifically, we first estimate a pair of initial intermediate flows using a high-resolution feature map for extracting local details. Then, we incorporate a sparse global matching branch to compen-sate for flow estimation, which consists of identifying flaws in initial flows and generating sparse flow compensation with a global receptive field. Finally, we adaptively merge the initial flow estimation with global flow compensation, yielding a more accurate intermediate flow. To evaluate the effectiveness of our method in handling large motion, we carefully curate a more challenging subset from commonly used benchmarks. Our method demonstrates the state-of-the-art performance on these VFI subsets with large motion.

We introduce X-Adapter, a universal upgrader to enable the pretrained plug-and-play modules (e.g., ControlNet, LoRA) to work directly with the upgraded text-to-image diffusion model (e.g., SDXL) without further retraining. We achieve this goal by training an additional network to control the frozen upgraded model with the new text-image data pairs. In detail, X-Adapter keeps a frozen copy of the old model to preserve the connectors of different plugins. Additionally, X-Adapter adds trainable mapping layers that bridge the decoders from models of different versions for feature remapping. The remapped features will be used as guidance for the upgraded model. To enhance the guidance ability of X-Adapter, we employ a null-text training strategy for the upgraded model. After training, we also introduce a two-stage denoising strategy to align the initial latents of X Adapter and the upgraded model. Thanks to our strategies, X-Adapter demonstrates universal compatibility with various plugins and also enables plugins of different versions to work together, thereby expanding the functionalities of diffusion community. To verify the effectiveness of the proposed method, we conduct extensive experiments and the results show that X-Adapter may facilitate wider application in the upgraded foundational diffusion model. Project page at: https://showlab.github.io/X-Adapter/.

We propose InstructDET, a data-centric method for referring object detection (ROD) that localizes target objects based on user instructions. While deriving from referring expressions (REC), the instructions we leverage are greatly diversified to encompass common user intentions related to object detection. For one image, we produce tremendous instructions that refer to every single object and different combinations of multiple objects. Each instruction and its corresponding object bounding boxes (bbxs) constitute one training data pair. In order to encompass common detection expressions, we involve emerging vision-language model (VLM) and large language model (LLM) to generate instructions guided by text prompts and object bbxs, as the generalizations of foundation models are effective to produce human-like expressions (e.g., describing object property, category, and relationship). We name our constructed dataset as InDET. It contains images, bbxs and generalized instructions that are from foundation models. Our InDET is developed from existing REC datasets and object detection datasets, with the expanding potential that any image with object bbxs can be incorporated through using our InstructDET method. By using our InDET dataset, we show that a conventional ROD model surpasses existing methods on standard REC datasets and our InDET test set. Our data-centric method InstructDET, with automatic data expansion by leveraging foundation models, directs a promising field that ROD can be greatly diversified to execute common object detection instructions.

Asynchronous action coordination presents a pervasive challenge in Multi-Agent Systems (MAS), which can be represented as a Stackelberg game (SG). However, the scalability of existing Multi-Agent Reinforcement Learning (MARL) methods based on SG is severely restricted by network architectures or environmental settings. To address this issue, we propose the Stackelberg Decision Transformer (STEER). It efficiently manages decision-making processes by incorporating the hierarchical decision structure of SG, the modeling capability of autoregressive sequence models, and the exploratory learning methodology of MARL. Our approach exhibits broad applicability across diverse task types and environmental configurations in MAS. Experimental results demonstrate both the convergence of our method towards Stackelberg equilibrium strategies and its superiority over strong baselines in complex scenarios.

Effective multi-intersection collaboration is pivotal for reinforcement-learning-based traffic signal control to alleviate congestion. Existing work mainly chooses neighboring intersections as collaborators. However, quite a lot of congestion, even some wide-range congestion, is caused by non-neighbors failing to collaborate. To address these issues, we propose to separate the collaborator selection as a second policy to be learned, concurrently being updated with the original signal-controlling policy. Specifically, the selection policy in real-time adaptively selects the best teammates according to phase- and intersection-level features. Empirical results on both synthetic and real-world datasets provide robust validation for the superiority of our approach, offering significant improvements over existing state-of-the-art methods. Code is available at https://github.com/bonaldli/CoSLight.

Traffic prediction, an essential component for intelligent transportation systems, endeavours to use historical data to foresee future traffic features at specific locations. Although existing traffic prediction models often emphasize developing complex neural network structures, their accuracy has not improved. Recently, large language models have shown outstanding capabilities in time series analysis. Differing from existing models, LLMs progress mainly through parameter expansion and extensive pretraining while maintaining their fundamental structures. Motivated by these developments, we propose a Spatial-Temporal Large Language Model (ST-LLM) for traffic prediction. In the ST-LLM, we define timesteps at each location as tokens and design a spatial-temporal embedding to learn the spatial location and global temporal patterns of these tokens. Additionally, we integrate these embeddings by a fusion convolution to each token for a unified spatial-temporal representation. Furthermore, we innovate a partially frozen attention strategy to adapt the LLM to capture global spatial-temporal dependencies for traffic prediction. Comprehensive experiments on real traffic datasets offer evidence that ST-LLM is a powerful spatial-temporal learner that outperforms state-of-the-art models. Notably, the ST-LLM also exhibits robust performance in both few-shot and zero-shot prediction scenarios. The code is publicly available at https://github.com/ChenxiLiu-HNU/ST-LLM.

Many few-shot segmentation (FSS) methods use cross attention to fuse support foreground (FG) into query features, regardless of the quadratic complexity. A recent advance Mamba can also well capture intra-sequence dependencies, yet the complexity is only linear. Hence, we aim to devise a cross (attention-like) Mamba to capture inter-sequence dependencies for FSS. A simple idea is to scan on support features to selectively compress them into the hidden state, which is then used as the initial hidden state to sequentially scan query features. Nevertheless, it suffers from (1) support forgetting issue: query features will also gradually be compressed when scanning on them, so the support features in hidden state keep reducing, and many query pixels cannot fuse sufficient support features; (2) intra-class gap issue: query FG is essentially more similar to itself rather than to support FG, i.e., query may prefer not to fuse support features but their own ones from the hidden state, yet the success of FSS relies on the effective use of support information. To tackle them, we design a hybrid Mamba network (HMNet), including (1) a support recapped Mamba to periodically recap the support features when scanning query, so the hidden state can always contain rich support information; (2) a query intercepted Mamba to forbid the mutual interactions among query pixels, and encourage them to fuse more support features from the hidden state. Consequently, the support information is better utilized, leading to better performance. Extensive experiments have been conducted on two public benchmarks, showing the superiority of HMNet. The code is available at https://github.com/Sam1224/HMNet.

Recent advancements in generation models have showcased remarkable capabilities in generating fantastic content. However, most of them are trained on proprietary high-quality data, and some models withhold their parameters and only provide accessible application programming interfaces (APIs), limiting their benefits for downstream tasks. To explore the feasibility of training a text-to-image generation model comparable to advanced models using publicly available resources, we introduce EvolveDirector. This framework interacts with advanced models through their public APIs to obtain text-image data pairs to train a base model. Our experiments with extensive data indicate that the model trained on generated data of the advanced model can approximate its generation capability. However, it requires large-scale samples of 10 million or more. This incurs significant expenses in time, computational resources, and especially the costs associated with calling fee-based APIs. To address this problem, we leverage pre-trained large vision-language models (VLMs) to guide the evolution of the base model. VLM continuously evaluates the base model during training and dynamically updates and refines the training dataset by the discrimination, expansion, deletion, and mutation operations. Experimental results show that this paradigm significantly reduces the required data volume. Furthermore, when approaching multiple advanced models, EvolveDirector can select the best samples generated by them to learn powerful and balanced abilities. The final trained model Edgen is demonstrated to outperform these advanced models. The code and model weights are available at https://github.com/showlab/EvolveDirector.

Inspired by the masked language modeling (MLM) in natural language processing tasks, the masked image modeling (MIM) has been recognized as a strong self-supervised pre-training method in computer vision. However, the high random mask ratio of MIM results in two serious problems: 1) the inadequate data utilization of images within each iteration brings prolonged pre-training, and 2) the high inconsistency of predictions results in unreliable generations, i.e., the prediction of the identical patch may be inconsistent in different mask rounds, leading to divergent semantics in the ultimately generated outcomes. To tackle these problems, we propose the efficient masked autoencoders with self-consistency (EMAE) to improve the pre-training efficiency and increase the consistency of MIM. In particular, we present a parallel mask strategy that divides the image into K non-overlapping parts, each of which is generated by a random mask with the same mask ratio. Then the MIM task is conducted parallelly on all parts in an iteration and the model minimizes the loss between the predictions and the masked patches. Besides, we design the self-consistency learning to further maintain the consistency of predictions of overlapping masked patches among parts. Overall, our method is able to exploit the data more efficiently and obtains reliable representations. Experiments on ImageNet show that EMAE achieves the best performance on ViT-Large with only 13% of MAE pre-training time using NVIDIA A100 GPUs. After pre-training on diverse datasets, EMAE consistently obtains state-of-the-art transfer ability on a variety of downstream tasks, such as image classification, object detection, and semantic segmentation.

Reinforcement learning (RL) can automatically learn a better policy through a trial-and-error paradigm and has been adopted to revolutionize and optimize traditional traffic signal control systems that are usually based on handcrafted methods. However, most existing RL-based models are either based on a single scenario or multiple independent scenarios, where each scenario has a separate simulation environment with predefined road network topology and traffic signal settings. These models implement training and testing in the same scenario, thus being strictly tied up with the specific setting and sacrificing model generalization heavily. While a few recent models could be trained by multiple scenarios, they require a huge amount of manual labor to label the intersection structure, hindering the model’s generalization. In this work, we aim at ageneralframework that could eliminate heavy labeling and model a variety of scenariossimultaneously. To this end, we propose a general Scenario-Agnostic (GESA) reinforcement learning framework for traffic signal control with: (1) A general plug-in module to map all different intersections into a unified structure, freeing us from the heavy manual labor to specify the structure of intersections; (2) A unified state and action space design to keep the model input and output consistently structured; (3) A large-scale co-training with multiple scenarios, leading to a generic traffic signal control algorithm. GESA can automatically handle various structured intersections from various cities without human labeling, and it co-trains a generalist agent to control traffic signals for multiple cities together, which also demonstrates superior transferability in zero-shot settings. In experiments, we demonstrate our algorithm as the first one that can be co-trained with seven different scenarios without manual annotation and gets13.27%higher rewards than baselines. When dealing with a new scenario, our model can still achieve9.39%higher rewards. The code, scenarios, and demos are available https://github.com/bonaldli/GESA.

Person clustering with multi-modal clues, including faces, bodies, and voices, is critical for various tasks, such as movie parsing and identity-based movie editing. Related methods such as multi-view clustering mainly project multi-modal features into a joint feature space. However, multi-modal clue features are usually rather weakly correlated due to the semantic gap from the modality-specific uniqueness. As a result, these methods are not suitable for person clustering. In this paper, we propose aRelation-AwareDistribution representation Network (RAD-Net) to generate adistribution representationfor multi-modal clues. The distribution representation of a clue is a vector consisting of the relation between this clue and all other clues from all modalities, thus beingmodality agnosticand good for person clustering. Accordingly, we introduce a graph-based method to construct distribution representation and employ a cyclic update policy to refine distribution representation progressively. Our method achieves substantial improvements of+6%and+8.2%in F-score on the Video Person-Clustering Dataset (VPCD) and VoxCeleb2 multi-view clustering dataset, respectively. Codes will be released athttps://github.com/bonaldli/RADNet.

Unsupervised domain adaptation (UDA) aims at adapting the model trained on a labeled source-domain dataset to an unlabeled target-domain dataset. The task of UDA on open-set person reidentification (re-ID) is even more challenging as the identities (classes) do not have overlap between the two domains. One major research direction was based on domain translation, which, however, has fallen out of favor in recent years due to inferior performance compared with pseudo-label-based methods. We argue that domain translation has great potential on exploiting valuable source-domain data but the existing methods did not provide proper regularization on the translation process. Specifically, previous methods only focus on maintaining the identities of the translated images while ignoring the intersample relations during translation. To tackle the challenges, we propose an end-to-end structured domain adaptation framework with an online relation-consistency regularization term. During training, the person feature encoder is optimized to model intersample relations on-the-fly for supervising relation-consistency domain translation, which in turn improves the encoder with informative translated images. The encoder can be further improved with pseudo labels, where the source-to-target translated images with ground-truth identities and target-domain images with pseudo identities are jointly used for training. In the experiments, our proposed framework is shown to achieve state-of-the-art performance on multiple UDA tasks of person re-ID. With the synthetic→real translated images from our structured domain-translation network, we achieved second place in the Visual Domain Adaptation Challenge (VisDA) in 2020.

Autoregressive language modeling (ALM) has been successfully used in self-supervised pre-training in Natural language processing (NLP). However, this paradigm has not achieved comparable results with other self-supervised approaches in computer vision (e.g., contrastive learning, masked image modeling). In this paper, we try to find the reason why autoregressive modeling does not work well on vision tasks. To tackle this problem, we fully analyze the limitation of visual autoregressive methods and proposed a novel stochastic autoregressive image modeling (named SAIM) by the two simple designs. First, we serialize the image into patches. Second, we employ the stochastic permutation strategy to generate an effective and robust image context which is critical for vision tasks. To realize this task, we create a parallel encoder-decoder training process in which the encoder serves a similar role to the standard vision transformer focusing on learning the whole contextual information, and meanwhile the decoder predicts the content of the current position so that the encoder and decoder can reinforce each other. Our method significantly improves the performance of autoregressive image modeling and achieves the best accuracy (83.9%) on the vanilla ViT-Base model among methods using only ImageNet-1K data. Transfer performance in downstream tasks also shows that our model achieves competitive performance. Code is available at https://github.com/qiy20/SAIM.

Uni6D is the first 6D pose estimation approach to employ a unified backbone network to extract features from both RGB and depth images. We discover that the principal reasons of Uni6D performance limitations are Instance-Outside and Instance-Inside noise. Uni6D’s simple pipeline design inherently introduces Instance-Outside noise from background pixels in the receptive field, while ignoring Instance-Inside noise in the input depth data. In this paper, we propose a two-step denoising approach for dealing with the aforementioned noise in Uni6D. To reduce noise from non-instance regions, an instance segmentation network is utilized in the first step to crop and mask the instance. A lightweight depth denoising module is proposed in the second step to calibrate the depth feature before feeding it into the pose regression network. Extensive experiments show that our Uni6Dv2 reliably and robustly eliminates noise, outperforming Uni6D without sacrificing too much inference efficiency. It also reduces the need for annotated real data that requires costly labeling.

Human-centric perceptions (e.g., pose estimation, human parsing, pedestrian detection, person re-identification, etc.) play a key role in industrial applications of visual models. While specific human-centric tasks have their own relevant semantic aspect to focus on, they also share the same underlying semantic structure of the human body. However, few works have attempted to exploit such homogeneity and design a general-propose model for human-centric tasks. In this work, we revisit a broad range of human-centric tasks and unify them in a minimalist manner. We propose UniHCP, a Unified Model for Human-Centric Perceptions, which unifies a wide range of human-centric tasks in a simplified end-to-end manner with the plain vision transformer architecture. With large-scale joint training on 33 human-centric datasets, UniHCP can outperform strong baselines on several in-domain and downstream tasks by direct evaluation. When adapted to a specific task, UniHCP achieves new SOTAs on a wide range of human-centric tasks, e.g., 69.8 mIoU on CIHP for human parsing, 86.18 mA on PA100K for attribute prediction, 90.3 mAP on Market1501 for ReID, and 85.8 JI on CrowdHuman for pedestrian detection, performing better than specialized models tailored for each task. The code and pretrained model are available at https://github.com/OpenGVLab/UniHCP.

Human-centric perceptions include a variety of vision tasks, which have widespread industrial applications, including surveillance, autonomous driving, and the metaverse. It is desirable to have a general pretrain model for versatile human-centric downstream tasks. This paper forges ahead along this path from the aspects of both benchmark and pretraining methods. Specifically, we propose a HumanBench based on existing datasets to comprehensively evaluate on the common ground the generalization abilities of different pretraining methods on 19 datasets from 6 diverse downstream tasks, including person ReID, pose estimation, human parsing, pedestrian attribute recognition, pedestrian detection, and crowd counting. To learn both coarse-grained and fine-grained knowledge in human bodies, we further propose a Projector AssisTed Hierarchical pretraining method (PATH) to learn diverse knowledge at different granularity levels. Comprehensive evaluations on HumanBench show that our PATH achieves new state-of-the-art results on 17 downstream datasets and on-par results on the other 2 datasets. The code will be publicly at https://github.com/OpenGVLab/HumanBench.

Semantic segmentation usually suffers from a long-tail data distribution. Due to the imbalanced number of samples across categories, the features of those tail classes may get squeezed into a narrow area in the feature space. Towards a balanced feature distribution, we introduce category-wise variation into the network predictions in the training phase such that an instance is no longer projected to a feature point, but a small region instead. Such a perturbation is highly dependent on the category scale, which appears as assigning smaller variation to head classes and larger variation to tail classes. In this way, we manage to close the gap between the feature areas of different categories, resulting in a more balanced representation. It is note-worthy that the introduced variation is discarded at the inference stage to facilitate a confident prediction. Although with an embarrassingly simple implementation, our method manifests itself in strong generalizability to various datasets and task settings. Extensive experiments suggest that our plug-in design lends itself well to a range of state-of-the-art approaches and boosts the performance on top of them.11Code: https://github.com/grantword8/BLV.

Open-vocabulary detection (OVD) is an object detection task aiming at detecting objects from novel categories beyond the base categories on which the detector is trained. Recent OVD methods rely on large-scale visual-language pre-trained models, such as CLIP, for recognizing novel objects. We identify the two core obstacles that need to be tackled when incorporating these models into detector training: (1) the distribution mismatch that happens when applying a VL-model trained on whole images to region recognition tasks; (2) the difficulty of localizing objects of unseen classes. To overcome these obstacles, we propose CORA, a DETR-style framework that adapts CLIP for Open-vocabulary detection by Region prompting and Anchor pre-matching. Region prompting mitigates the whole-to-region distribution gap by prompting the region features of the CLIP-based region classifier. Anchor pre-matching helps learning generalizable object localization by a class-aware matching mechanism. We evaluate CORA on the COCO OVD benchmark, where we achieve 41.7 AP50 on novel classes, which outperforms the previous SOTA by 2.4 AP50 even without resorting to extra training data. When extra training data is available, we train CORA+ on both ground-truth base-category annotations and additional pseudo bounding box labels computed by CORA. CORA+ achieves 43.1 AP50 on the COCO OVD benchmark and 28.1 box APr on the LVIS OVD benchmark. The code is available at https://github.com/tgxs002/CORA.

Video-text retrieval has greatly benefited from the massive web video in recent years, while the performance is still limited to the weak supervision from the uncurated data. In this work, we propose to leverage the well-represented information of each original modality and exploit complementary information in two views of the same video, i.e., video clips and captions, by using one view to obtain positive samples with the neighboring samples of the other. Respecting the hierarchical organization of real-world data, we further design a hierarchical cross-modal pre-training method (HCP) to learn good representations in the common embedding space. We evaluate the pre-trained model on three downstream tasks, i.e. text-to-video retrieval, action step localization and video question answering and our method outperforms previous works under the same setting.

Recent years have witnessed a rapid growth of deep generative models, with text-to-image models gaining significant attention from the public. However, existing models often generate images that do not align well with human preferences, such as awkward combinations of limbs and facial expressions. To address this issue, we collect a dataset of human choices on generated images from the Stable Foundation Discord channel. Our experiments demonstrate that current evaluation metrics for generative models do not correlate well with human choices. Thus, we train a human preference classifier with the collected dataset and derive a Human Preference Score (HPS) based on the classifier. Using HPS, we propose a simple yet effective method to adapt Stable Diffusion to better align with human preferences. Our experiments show that HPS outperforms CLIP in predicting human choices and has good generalization capability toward images generated from other models. By tuning Stable Diffusion with the guidance of HPS, the adapted model is able to generate images that are more preferred by human users. The project page is available here: https://tgxs002.github.io/alignsd-web/.

Referring Expression Segmentation (RES) is a widely explored multi-modal task, which endeavors to segment the pre-existing object within a single image with a given linguistic expression. However, in broader real-world scenarios, it is not always possible to determine if the described object exists in a specific image. Generally, a collection of images is available, some of which potentially contain the target objects. To this end, we propose a more realistic setting, named Group-wise Referring Expression Segmentation (GRES), which expands RES to a group of related images, allowing the described objects to exist in a subset of the input image group. To support this new setting, we introduce an elaborately compiled dataset named Grouped Referring Dataset (GRD), containing complete group-wise annotations of the target objects described by given expressions. Moreover, we also present a baseline method named Grouped Referring Segmenter (GRSer), which explicitly captures the language-vision and intra-group vision-vision interactions to achieve state-of-the-art results on the proposed GRES setting and related tasks, such as Co-Salient Object Detection and traditional RES. Our dataset and codes are publicly released in https://github.com/shikras/d-cube.

Masked Autoencoders (MAE) and its variants have proven to be effective for pretraining large-scale Vision Transformers (ViTs). However, small-scale models do not benefit from the pretraining mechanisms due to limited capacity. Sparse training is a method of transferring representations from large models to small ones by pruning unimportant parameters. However, naively combining MAE finetuning with sparse training make the network task-specific, resulting in the loss of task-agnostic knowledge, which is crucial for model generalization. In this paper, we aim to reduce model complexity from large vision transformers pretrained by MAE with assistant of sparse training. We summarize various sparse training methods to prune large vision transformers during MAE pretraining and finetuning stages, and discuss their shortcomings. To improve learning both task-agnostic and task-specific knowledge, we propose SparseMAE, a novel two-stage sparse training method that includes sparse pretraining and sparse finetuning. In sparse pretraining, we dynamically prune a small-scale sub-network from a ViT-Base. During finetuning, the sparse sub-network adaptively changes its topology connections under the task-agnostic knowledge of the full model. Extensive experimental results demonstrate the effectiveness of our method and its superiority on small-scale vision transformers. Code will be available at https://github.com/aojunzz/SparseMAE.

This paper proposes to learn Multi-task, Multi-modal Direct Acyclic Graphs (MM-DAGs), which are commonly observed in complex systems, e.g., traffic, manufacturing, and weather systems, whose variables are multi-modal with scalars, vectors, and functions. This paper takes the traffic congestion analysis as a concrete case, where a traffic intersection is usually regarded as a DAG. In a road network of multiple intersections, different intersections can only have someoverlapping and distinct variables observed. For example, a signalized intersection has traffic light-related variables, whereas unsignalized ones do not. This encourages the multi-task design: with each DAG as a task, the MM-DAG tries to learn the multiple DAGs jointly so that their consensus and consistency are maximized. To this end, we innovatively propose a multi-modal regression for linear causal relationship description of different variables. Then we develop a novel Causality Difference (CD) measure and its differentiable approximator. Compared with existing SOTA measures, CD can penalize the causal structural difference among DAGs with distinct nodes and can better consider the uncertainty of causal orders. We rigidly prove our design's topological interpretation and consistency properties. We conduct thorough simulations and one case study to show the effectiveness of our MM-DAG. The code is available under https://github.com/Lantian72/MM-DAG.

Public large-scale text-to-image diffusion models, such as Stable Diffusion, have gained significant attention from the community. These models can be easily customized for new concepts using low-rank adaptations (LoRAs). However, the utilization of multiple-concept LoRAs to jointly support multiple customized concepts presents a challenge. We refer to this scenario as decentralized multi-concept customization, which involves single-client concept tuning and center-node concept fusion. In this paper, we propose a new framework called Mix-of-Show that addresses the challenges of decentralized multi-concept customization, including concept conflicts resulting from existing single-client LoRA tuning and identity loss during model fusion. Mix-of-Show adopts an embedding-decomposed LoRA (ED-LoRA) for single-client tuning and gradient fusion for the center node to preserve the in-domain essence of single concepts and support theoretically limitless concept fusion. Additionally, we introduce regionally controllable sampling, which extends spatially controllable sampling (e.g., ControlNet and T2I-Adapter) to address attribute binding and missing object problems in multi-concept sampling. Extensive experiments demonstrate that Mix-of-Show is capable of composing multiple customized concepts with high fidelity, including characters, objects, and scenes.

Current deep networks are very data-hungry and benefit from training on large-scale datasets, which are often time-consuming to collect and annotate. By contrast, synthetic data can be generated infinitely using generative models such as DALL-E and diffusion models, with minimal effort and cost. In this paper, we present DatasetDM, a generic dataset generation model that can produce diverse synthetic images and the corresponding high-quality perception annotations (e.g., segmentation masks, and depth). Our method builds upon the pre-trained diffusion model and extends text-guided image synthesis to perception data generation. We show that the rich latent code of the diffusion model can be effectively decoded as accurate perception annotations using a decoder module. Training the decoder only needs less than 1% (around 100 images) of manually labeled images, enabling the generation of an infinitely large annotated dataset. Then these synthetic data can be used for training various perception models on downstream tasks. To showcase the power of the proposed approach, we generate datasets with rich dense pixel-wise labels for a wide range of downstream tasks, including semantic15 segmentation, instance segmentation, and depth estimation. Notably, it achieves 1) state-of-the-art results on semantic segmentation and instance segmentation; 2) significantly more efficient and robust in domain generalization than the real data; 3) state-of-the-art results in zero-shot segmentation setting; and 4) flexibility for efficient application and novel task composition (e.g., image editing)

Detecting objects based on language information is a popular task that includes Open-Vocabulary object Detection (OVD) and Referring Expression Comprehension (REC). In this paper, we advance them to a more practical setting called *Described Object Detection* (DOD) by expanding category names to flexible language expressions for OVD and overcoming the limitation of REC only grounding the pre-existing object. We establish the research foundation for DOD by constructing a *Description Detection Dataset* ($D^3$). This dataset features flexible language expressions, whether short category names or long descriptions, and annotating all described objects on all images without omission. By evaluating previous SOTA methods on $D^3$, we find some troublemakers that fail current REC, OVD, and bi-functional methods. REC methods struggle with confidence scores, rejecting negative instances, and multi-target scenarios, while OVD methods face constraints with long and complex descriptions. Recent bi-functional methods also do not work well on DOD due to their separated training procedures and inference strategies for REC and OVD tasks. Building upon the aforementioned findings, we propose a baseline that largely improves REC methods by reconstructing the training data and introducing a binary classification sub-task, outperforming existing methods. Data and code are available at https://github.com/shikras/d-cube and related works are tracked in https://github.com/Charles-Xie/awesome-described-object-detection.

Recent years person re-identification (ReID) has been developed rapidly due to its broad practical applications. Most existing benchmarks assume that the same person wears the same clothes across captured images, while, in real-world scenarios, person may change his/her clothes frequently. Thus the Clothes-Changing person ReID (CC-ReID) problem is introduced and several related benchmarks are established. CC-ReID is a very difficult task as the main visual characteristics of a human body, clothes, are different between query and gallery, and clothes-irrelevant features are relatively weak. To promote the research and applications of person ReID in clothes-changing scenarios, in this paper, we introduce a new task called Clothes Template based Clothes-Changing person ReID (CTCC-ReID), where the query image is enhanced by a clothes template which shares similar visual patterns with the clothes of the target person image in the gallery. So, ReID methods are encouraged to jointly consider the original query image and the given clothes template for retrieval in the proposed CTCC-ReID setting. To facilitate research works on CTCC-ReID, we construct a novel large-scale ReID dataset named ClOthes ChAnging person Set Plus (COCAS+), which contains both realistic and synthetic clothes-changing person images with manually collected clothes templates. Furthermore, we propose a novel Dual-Attention Biometric-Clothes Transfusion Network (DualBCT-Net) for CTCC-ReID, which can effectively learn to extract biometric features from the original query person image and clothes features from the given clothes template and then fuse them through a Dual-Attention Fusion Module. Extensive experimental results show that the proposed CTCC-ReID setting and COCAS+ dataset can help greatly push the performance of clothes-changing ReID toward practical applications, and synthetic data is impressively effective for CTCC-ReID. What’s more, the proposed DualBCT-Net shows significant improvements over state-of-the-art methods on the CTCC-ReID task. COCAS+ and code of DualBCT-Net will be released inhttps://github.com/Chenhaobin/COCAS-plus.

Recent adversarial attack works attempt to improve the transferability by applying various differentiable transformations on input images. Considering the differentiable transformations and the original model together as a new model, these methods can be regarded as model augmentation that effectively derives an ensemble of models from the single original model. Despite their impressive performance, the model augmentation policies used in these methods are manually designed by experimental attempts, leaving the design of model augmentation policy an open question. In this paper, we propose an Automatic Model Augmentation (AutoMA) approach to find a strong model augmentation policy for transferable adversarial attacks. Specifically, we design a discrete search space that contains various diffierentiable transformations with different parameters and adopt reinforcement learning to search for the strong augmentation policy. The sampled augmentation policies together with the rewards they obtain during the searching process reveal several valuable observations for designing more powerful attacks using model augmentation policy:1) Augmentation transformations on color space are less effective; 2) The transformation type diversity matters; and 3) Using small distortion for geometric transformations while larger distortion for intensity transformations.Extensive experiments show that the augmentation policy found by AutoMA achieves superior performance than existing manually designed policies in a wide range of cases.

Road network and trajectory representation learning are essential for traffic systems since the learned representation can be directly used in various downstream tasks (e.g., traffic speed inference, travel time estimation). However, most existing methods only contrast within the same scale, i.e., treating road network and trajectory separately, which ignores valuable inter-relations. In this paper, we aim to propose a unified framework that jointly learns the road network and trajectory representations end-to-end. We design domain-specific augmentations for road-road contrast and trajectory-trajectory contrast separately, i.e., road segment with its contextual neighbors and trajectory with its detour replaced and dropped alternatives, respectively. On top of that, we further introduce the road-trajectory cross-scale contrast to bridge the two scales by maximizing the total mutual information. Unlike the existing cross-scale contrastive learning methods on graphs that only contrast a graph and its belonging nodes, the contrast between road segment and trajectory is elaborately tailored via novel positive sampling and adaptive weighting strategies. We conduct prudent experiments based on two real-world datasets with four downstream tasks, demonstrating improved performance and effectiveness.

As RGB-D sensors become more affordable, using RGB- D images to obtain high-accuracy 6D pose estimation results becomes a better option. State-of-the-art approaches typically use different backbones to extract features for RGB and depth images. They use a 2D CNN for RGB images and a perpixel point cloud network for depth data, as well as a fusion network for feature fusion. We find that the essential reason for using two independent backbones is the “projection breakdown” problem. In the depth image plane, the projected 3D structure of the physical world is preserved by the 1D depth value and its built-in 2D pixel coordinate (UV). Any spatial transformation that modifies UV, such as resize, flip, crop, or pooling operations in the CNN pipeline, breaks the binding between the pixel value and UV coordinate. As a consequence, the 3D structure is no longer preserved by a modified depth image or feature. To address this issue, we propose a simple yet effective method denoted as Uni6D that explicitly takes the extra UV data along with RGB-D images as input. Our method has a Unified CNN framework for 6D pose estimation with a single CNN backbone. In particular, the architecture of our method is based on Mask R-CNN with two extra heads, one named RT head for directly predicting 6D pose and the other named abc head for guiding the network to map the visible points to their coordinates in the 3D model as an auxiliary module. This end-to-end approach balances simplicity and accuracy, achieving comparable accuracy with state of the arts and 7.2x faster inference speed on the YCB-Video dataset.

Self-supervised learning (SSL) holds promise in leveraging large amounts of unlabeled data. However, the success of popular SSL methods has limited on single-centric-object images like those in ImageNet and ignores the correlation among the scene and instances, as well as the semantic difference of instances in the scene. To address the above problems, we propose a Unified Self-supervised Visual Pre-training (UniVIP), a novel self-supervised framework to learn versatile visual representations on either single-centric-object or non-iconic dataset. The framework takes into account the representation learning at three levels: 1) the similarity of scene-scene, 2) the correlation of scene-instance, 3) the discrimination of instance-instance. During the learning, we adopt the optimal transport algorithm to automatically measure the discrimination of instances. Massive experiments show that Uni-VIP pre-trained on non-iconic COCO achieves state-of-the-art transfer performance on a variety of downstream tasks, such as image classification, semi-supervised learning, object detection and segmentation. Furthermore, our method can also exploit single-centric-object dataset such as ImageNet and outperforms BYOL by 2.5% with the same pre-training epochs in linear probing, and surpass current self-supervised object detection methods on COCO dataset, demonstrating its universality and potential.

The pretrain-finetune paradigm is a classical pipeline in visual learning. Recent progress on unsupervised pretraining methods shows superior transfer performance to their supervised counterparts. This paper revisits this phenomenon and sheds new light on understanding the transferability gap between unsupervised and supervised pretraining from a multilayer perceptron (MLP) perspective. While previous works [6], [8], [17] focus on the effectiveness of MLP on unsupervised image classification where pretraining and evaluation are conducted on the same dataset, we reveal that the MLP projector is also the key factor to better transferability of unsupervised pretraining methods than supervised pretraining methods. Based on this observation, we attempt to close the transferability gap between supervised and unsupervised pretraining by adding an MLP projector before the classifier in supervised pretraining. Our analysis indicates that the MLP projector can help retain intra-class variation of visual features, decrease the feature distribution distance between pretraining and evaluation datasets, and reduce feature redundancy. Extensive experiments on public benchmarks demonstrate that the added MLP projector significantly boosts the transferability of supervised pretraining, e.g. +7.2% top-1 accuracy on the concept generalization task, +5.8% top-1 accuracy for linear evaluation on 12 -domain classification tasks, and +0.8% AP on COCO object detection task, making supervised pretraining comparable or even better than unsupervised pretraining.

The crux of semi-supervised semantic segmentation is to assign adequate pseudo-labels to the pixels of unlabeled images. A common practice is to select the highly confident predictions as the pseudo ground-truth, but it leads to a problem that most pixels may be left unused due to their unreliability. We argue that every pixel matters to the model training, even its prediction is ambiguous. Intuitively, an unreliable prediction may get confused among the top classes (i.e., those with the highest probabilities), however, it should be confident about the pixel not belonging to the remaining classes. Hence, such a pixel can be convincingly treated as a negative sample to those most unlikely categories. Based on this insight, we develop an effective pipeline to make sufficient use of unlabeled data. Concretely, we separate reliable and unreliable pixels via the entropy of predictions, push each unreliable pixel to a category-wise queue that consists of negative samples, and manage to train the model with all candidate pixels. Considering the training evolution, where the prediction becomes more and more accurate, we adaptively adjust the threshold for the reliable-unreliable partition. Experimental results on various benchmarks and training settings demonstrate the superiority of our approach over the state-of-the-art alternatives.11Project: https://haochen-wang409.github.io/U2PL.

Occluded person re-identification (ReID) aims at matching occluded person images to holistic ones across different camera views. Target Pedestrians (TP) are often disturbed by Non-Pedestrian Occlusions (NPO) and Non-Target Pedestrians (NTP). Previous methods mainly focus on increasing the model's robustness against NPO while ignoring feature contamination from NTP. In this paper, we propose a novel Feature Erasing and Diffusion Network (FED) to simultaneously handle challenges from NPO and NTP. Specifically, aided by the NPO augmentation strategy that simulates NPO on holistic pedestrian images and gen-erates precise occlusion masks, NPO features are explicitly eliminated by our proposed Occlusion Erasing Module (OEM). Subsequently, we diffuse the pedestrian representations with other memorized features to synthesize the NTP characteristics in the feature space through the novel Feature Diffusion Module (FDM). With the guidance of the occlusion scores from OEM, the feature diffusion process is conducted on visible body parts, thereby improving the quality of the synthesized NTP characteristics. We can greatly improve the model's perception ability towards TP and alleviate the influence of NPO and NTP by jointly optimizing OEM and FDM. Furthermore, the proposed FDM works as an auxiliary module for training and will not be engaged in the inference phase, thus with high flexibility. Experiments on occluded and holistic person ReID benchmarks demonstrate the superiority of FED over state-of-the-art methods.

Existing symmetric contrastive learning methods suffer from collapses (complete and dimensional) or quadratic complexity of objectives. Departure from these methods which maximize mutual information of two generated views, along either instance or feature dimension, the proposed paradigm introduces intermediate variables at the feature level, and maximizes the consistency between variables and representations of each view. Specifically, the proposed intermediate variables are the nearest group of base vectors to representations. Hence, we call the proposed method ARB (Align Representations with Base). Compared with other symmetric approaches, ARB 1) does not require negative pairs, which leads the complexity of the overall objective function is in linear order, 2) reduces feature redundancy, increasing the information density of training samples, 3) is more robust to output dimension size, which out-performs previous feature-wise arts over 28% Top-1 accuracy on ImageNet-100under low-dimension settings.

The essence of unsupervised anomaly detection is to learn the compact distribution of normal samples and detect outliers as anomalies in testing. Meanwhile, the anomalies in real-world are usually subtle and fine-grained in a high-resolution image especially for industrial applications. Towards this end, we propose a novel framework for unsupervised anomaly detection and localization. Our method aims at learning dense and compact distribution from normal images with a coarse-to-fine alignment process. The coarse alignment stage standardizes the pixel-wise position of objects in both image and feature levels. The fine alignment stage then densely maximizes the similarity of features among all corresponding locations in a batch. To facilitate the learning with only normal images, we propose a new pretext task called non-contrastive learning for the fine alignment stage. Non-contrastive learning extracts robust and discriminating normal image representations without making assumptions on abnormal samples, and it thus empowers our model to generalize to various anomalous scenarios. Extensive experiments on two typical industrial datasets of MVTec AD and BenTech AD demonstrate that our framework is effective in detecting various real-world defects and achieves a new state-of-the-art in industrial unsupervised anomaly detection.

Visual tasks vary a lot in their output formats and concerned contents, therefore it is hard to process them with an identical structure. One main obstacle lies in the high-dimensional outputs in object-level visual tasks. In this paper, we propose an object-centric vision framework, Obj2Seq. Obj2Seq takes objects as basic units, and regards most object-level visual tasks as sequence generation problems of objects. Therefore, these visual tasks can be decoupled into two steps. First recognize objects of given categories, and then generate a sequence for each of these objects. The definition of the output sequences varies for different tasks, and the model is supervised by matching these sequences with ground-truth targets. Obj2Seq is able to flexibly determine input categories to satisfy customized requirements, and be easily extended to different visual tasks. When experimenting on MS COCO, Obj2Seq achieves 45.7% AP on object detection, 89.0% AP on multi-label classification and 65.0% AP on human pose estimation. These results demonstrate its potential to be generally applied to different visual tasks. Code has been made available at: https://github.com/CASIA-IVA-Lab/Obj2Seq.

Self-training has shown great potential in semi-supervised learning. Its core idea is to use the model learned on labeled data to generate pseudo-labels for unlabeled samples, and in turn teach itself. To obtain valid supervision, active attempts typically employ a momentum teacher for pseudo-label prediction yet observe the confirmation bias issue, where the incorrect predictions may provide wrong supervision signals and get accumulated in the training process. The primary cause of such a drawback is that the prevailing self-training framework acts as guiding the current state with previous knowledge because the teacher is updated with the past student only. To alleviate this problem, we propose a novel self-training strategy, which allows the model to learn from the future. Concretely, at each training step, we first virtually optimize the student (i.e., caching the gradients without applying them to the model weights), then update the teacher with the virtual future student, and finally ask the teacher to produce pseudo-labels for the current student as the guidance. In this way, we manage to improve the quality of pseudo-labels and thus boost the performance. We also develop two variants of our future-self-training (FST) framework through peeping at the future both deeply (FST-D) and widely (FST-W). Taking the tasks of unsupervised domain adaptive semantic segmentation and semi-supervised semantic segmentation as the instances, we experimentally demonstrate the effectiveness and superiority of our approach under a wide range of settings. Code is available at https://github.com/usr922/FST.

Unsupervised pretraining methods for object detection aim to learn object discrimination and localization ability from large amounts of images. Typically, recent works design pretext tasks that supervise the detector to predict the defined object priors. They normally leverage heuristic methods to produce object priors, \emph{e.g.,} selective search, which separates the prior generation and detector learning and leads to sub-optimal solutions. In this work, we propose a novel object detection pretraining framework that could generate object priors and learn detectors jointly by generating accurate object priors from the model itself. Specifically, region priors are extracted by attention maps from the encoder, which highlights foregrounds. Instance priors are the selected high-quality output bounding boxes of the detection decoder. By assuming objects as instances in the foreground, we can generate object priors with both region and instance priors. Moreover, our object priors are jointly refined along with the detector optimization. With better object priors as supervision, the model could achieve better detection capability, which in turn promotes the object priors generation. Our method improves the competitive approaches by \textbf{+1.3 AP}, \textbf{+1.7 AP} in 1\% and 10\% COCO low-data regimes object detection.

Correspondence pruning aims to correctly remove false matches (outliers) from an initial set of putative correspondences. The pruning process is challenging since putative matches are typically extremely unbalanced, largely dominated by outliers, and the random distribution of such outliers further complicates the learning process for learning-based methods. To address this issue, we propose to progressively prune the correspondences via a local-to-global consensus learning procedure. We introduce a "pruning" block that lets us identify reliable candidates among the initial matches according to consensus scores estimated using local-to-global dynamic graphs. We then achieve progressive pruning by stacking multiple pruning blocks sequentially. Our method outperforms state-of-the-arts on robust line fitting, camera pose estimation and retrieval-based image localization benchmarks by significant margins and shows promising generalization ability to different datasets and detector/descriptor combinations.

Exploiting the relationships between attributes is a key challenge for improving multiple facial attribute recognition. In this work, we are concerned with two types of correlations that are spatial and non-spatial relationships. For the spatial correlation, we aggregate attributes with spatial similarity into a part-based group and then introduce a Group Attention Learning to generate the group attention and the part-based group feature. On the other hand, to discover the non-spatial relationship, we model a group-based Graph Correlation Learning to explore affinities of predefined part-based groups. We utilize such affinity information to control the communication between all groups and then refine the learned group features. Overall, we propose a unified network called Multi-scale Group and Graph Network. It incorporates these two newly proposed learning strategies and produces coarse-to-fine graph-based group features for improving facial attribute recognition. Comprehensive experiments demonstrate that our approach outperforms the state-of-the-art methods.

Different from close-set classification task, deep face recognition models are often used in open-set scenarios, where the models need to handle arbitrary faces. Open-set adversarial attacks can identify the vulnerability of deep face recognition models. Compared to time-consuming iterative gradient-based methods, generator-based methods can produce adversarial examples with only one forward pass, which greatly improves attack efficiency. However, existing generator-based attack methods need to train an individual model for each target identity and can only generate a fixed perturbation pattern regardless of different attack intensity constraints, which is impractical and sub-optimal for open-set adversarial attacks. In this paper, we propose an efficient generator-based Single Model ARbitrary Target (SMART) approach for open-set adversarial attacks against deep face recognition models. Given an arbitrary source-target face image pair, SMART first generates an additive perturbation and then adds it to the source image to obtain the final adversarial face image. After the training with various source-target pairs randomly sampled on large scale face images, SMART could effectively learn inherent perturbation patterns for arbitrary source-target face images pairs. Besides, we also propose a novel Constraint-aware Adversarial Decoder (CAD) module, which makes SMART the first generator-based method that could produce adaptive adversarial patterns according to different constraints on attack intensity. Extensive experimental results in various settings demonstrate the effectiveness of the proposed method.

Transformer has been widely used for self-supervised pre-training in Natural Language Processing (NLP) and achieved great success. However, it has not been fully explored in visual self-supervised learning. Meanwhile, previous methods only consider the high-level feature and learning representation from a global perspective, which may fail to transfer to the downstream dense prediction tasks focusing on local features. In this paper, we present a novel Masked Self-supervised Transformer approach named MST, which can explicitly capture the local context of an image while preserving the global semantic information. Specifically, inspired by the Masked Language Modeling (MLM) in NLP, we propose a masked token strategy based on the multi-head self-attention map, which dynamically masks some tokens of local patches without damaging the crucial structure for self-supervised learning. More importantly, the masked tokens together with the remaining tokens are further recovered by a global image decoder, which preserves the spatial information of the image and is more friendly to the downstream dense prediction tasks. The experiments on multiple datasets demonstrate the effectiveness and generality of the proposed method. For instance, MST achieves Top-1 accuracy of 76.9% with DeiT-S only using 300-epoch pre-training by linear evaluation, which outperforms supervised methods with the same epoch by 0.4% and its comparable variant DINO by 1.0%. For dense prediction tasks, MST also achieves 42.7% mAP on MS COCO object detection and 74.04% mIoU on Cityscapes segmentation only with 100-epoch pre-training.

With the explosion of digital data in recent years, continuously learning new tasks from a stream of data without forgetting previously acquired knowledge has become increasingly important. In this paper, we propose a new continual learning (CL) setting, namely "continual representation learning", which focuses on learning better representation in a continuous way. We also provide two large-scale multi-step benchmarks for biometric identification, where the visual appearance of different classes are highly relevant. In contrast to requiring the model to recognize more learned classes, we aim to learn feature representation that can be better generalized to not only previously unseen images but also unseen classes/identities. For the new setting, we propose a novel approach that performs the knowledge distillation over a large number of identities by applying the neighbourhood selection and consistency relaxation strategies to improve scalability and flexibility of the continual learning model. We demonstrate that existing CL methods can improve the representation in the new setting, and our method achieves better results than the competitors.

Clustering has many applications in research and industry. However, traditional clustering methods, such as K-means, DBSCAN and HAC, impose oversimplifying assumptions and thus are not well-suited to face clustering. To adapt to the distribution of realistic problems, a natural approach is to use Graph Convolutional Networks (GCNs) to enhance features for clustering. However, GCNs can only utilize local information, which ignores the overall characterisitcs of the clusters. In this paper, we propose a Density-Aware Feature Embedding Network (DA-Net) for the task of face clustering, which utilizes both local and non-local information, to learn a robust feature embedding. Specifically, DA-Net uses GCNs to aggregate features locally, and then incorporates non-local information using a density chain, which is a chain of faces from low density to high density. This density chain exploits the non-uniform distribution of face images in the dataset. Then, an LSTM takes the density chain as input to generate the final feature embedding. Once this embedding is generated, traditional clustering methods, such as density-based clustering, can be used to obtain the final clustering results. Extensive experiments verify the effectiveness of the proposed feature embedding method, which can achieve state-of-the-art performance on public benchmarks.

Face clustering is an essential tool for exploiting the unlabeled face data, and has a wide range of applications including face annotation and retrieval. Recent works show that supervised clustering can result in noticeable performance gain. However, they usually involve heuristic steps and require numerous overlapped subgraphs, severely restricting their accuracy and efficiency. In this paper, we propose a fully learnable clustering framework without requiring a large number of overlapped subgraphs. Instead, we transform the clustering problem into two sub-problems. Specifically, two graph convolutional networks, named GCN-V and GCN-E, are designed to estimate the confidence of vertices and the connectivity of edges, respectively. With the vertex confidence and edge connectivity, we can naturally organize more relevant vertices on the affinity graph and group them into clusters. Experiments on two large-scale benchmarks show that our method significantly improves clustering accuracy and thus performance of the recognition models trained on top, yet it is an order of magnitude more efficient than existing supervised methods.

Recent years have witnessed great progress in person re-identification (re-id). Several academic benchmarks such as Market1501, CUHK03 and DukeMTMC play important roles to promote the re-id research. To our best knowledge, all the existing benchmarks assume the same person will have the same clothes. While in real-world scenarios, it is very often for a person to change clothes. To address the clothes changing person re-id problem, we construct a novel large-scale re-id benchmark named Clothes Changing Person Set (COCAS), which provides multiple images of the same identity with different clothes. COCAS totally contains 62,382 body images from 5,266 persons. Based on COCAS, we introduce a new person re-id setting for clothes changing problem, where the query includes both a clothes template and a person image taking another clothes. Moreover, we propose a two-branch network named Biometric-Clothes Network (BC-Net) which can effectively integrate biometric and clothes feature for re-id under our setting. Experiments show that it is feasible for clothes changing re-id with clothes templates.

Existing dominant approaches for cross-modal video-text retrieval task are to learn ajoint embedding space to measure the cross-modal similarity. However, these methods rarely explore long-range dependency inside video frames or textual words leading to insufficient textual and visual details. In this paper, we propose a stacked convolutional deep encoding network for video-text retrieval task, which considers to simultaneously encode long-range and short-range dependency in the videos and texts. Specifically, a multi-scale dilated convolutional (MSDC) block within our approach is able to encode short-range temporal cues between video frames or text words by adopting different scales of kernel size and dilation size of convolutional layer. A stacked structure is designed to expand the receptive fields by repeatedly adopting the MSD- C block, which further captures the long-range relations between these cues. Moreover, to obtain more robust textual representations, we fully utilize the powerful language model named Transformer in two stages: pretraining phrase and fine-tuning phrase. Extensive experiments on two different benchmark datasets (MSR-VTT, MSVD) show that our proposed method outperforms other state-of-the-art approaches.

Domain adaptive object re-ID aims to transfer the learned knowledge from the labeled source domain to the unlabeled target domain to tackle the open-class re-identification problems. Although state-of-the-art pseudo-label-based methods have achieved great success, they did not make full use of all valuable information because of the domain gap and unsatisfying clustering performance. To solve these problems, we propose a novel self-paced contrastive learning framework with hybrid memory. The hybrid memory dynamically generates source-domain class-level, target-domain cluster-level and un-clustered instance-level supervisory signals for learning feature representations. Different from the conventional contrastive learning strategy, the proposed framework jointly distinguishes source-domain classes, and target-domain clusters and un-clustered instances. Most importantly, the proposed self-paced method gradually creates more reliable clusters to refine the hybrid memory and learning targets, and is shown to be the key to our outstanding performance. Our method outperforms state-of-the-arts on multiple domain adaptation tasks of object re-ID and even boosts the performance on the source domain without any extra annotations. Our generalized version on unsupervised object re-ID surpasses state-of-the-art algorithms by considerable 16.7% and 7.9% on Market-1501 and MSMT17 benchmarks.

Cosine-based softmax losses significantly improve the performance of deep face recognition networks. However, these losses always include sensitive hyper-parameters which can make training process unstable, and it is very tricky to set suitable hyper parameters for a specific dataset. This paper addresses this challenge by directly designing the gradients for training in an adaptive manner. We first investigate and unify previous cosine softmax losses from the perspective of gradients. This unified view inspires us to propose a novel gradient called P2SGrad (Probability-to-Similarity Gradient), which leverages a cosine similarity instead of classification probability to control the gradients for updating neural network parameters. P2SGrad is adaptive and hyper-parameter free, which makes training process more efficient and faster. We evaluate our P2SGrad on three face recognition benchmarks, LFW, MegaFace, and IJB-C. The results show that P2SGrad is stable in training, robust to noise, and achieves state-of-the-art performance on all the three benchmarks.

The cosine-based softmax losses and their variants achieve great success in deep learning based face recognition. However, hyperparameter settings in these losses have significant influences on the optimization path as well as the final recognition performance. Manually tuning those hyperparameters heavily relies on user experience and requires many training tricks. In this paper, we investigate in depth the effects of two important hyperparameters of cosine-based softmax losses, the scale parameter and angular margin parameter, by analyzing how they modulate the predicted classification probability. Based on these analysis, we propose a novel cosine-based softmax loss, AdaCos, which is hyperparameter-free and leverages an adaptive scale parameter to automatically strengthen the training supervisions during the training process. We apply the proposed AdaCos loss to large-scale face verification and identification datasets, including LFW, MegaFace, and IJB-C 1:1 Verification. Our results show that training deep neural networks with the AdaCos loss is stable and able to achieve high face recognition accuracy. Our method outperforms state-of-the-art softmax losses on all the three datasets.

Learning discriminative image feature embeddings is of great importance to visual recognition. To achieve better feature embeddings, most current methods focus on designing different network structures or loss functions, and the estimated feature embeddings are usually only related to the input images. In this paper, we propose Memory-based Neighbourhood Embedding (MNE) to enhance a general CNN feature by considering its neighbourhood. The method aims to solve two critical problems, i.e., how to acquire more relevant neighbours in the network training and how to aggregate the neighbourhood information for a more discriminative embedding. We first augment an episodic memory module into the network, which can provide more relevant neighbours for both training and testing. Then the neighbours are organized in a tree graph with the target instance as the root node. The neighbourhood information is gradually aggregated to the root node in a bottom-up manner, and aggregation weights are supervised by the class relationships between the nodes. We apply MNE on image search and few shot learning tasks. Extensive ablation studies demonstrate the effectiveness of each component, and our method significantly outperforms the state-of-the-art approaches.

Person re-identification (ReID) is to identify pedestrians observed from different camera views based on visual appearance. It is a challenging task due to large pose variations, complex background clutters and severe occlusions. Recently, human pose estimation by predicting joint locations was largely improved in accuracy. It is reasonable to use pose estimation results for handling pose variations and background clutters, and such attempts have obtained great improvement in ReID performance. However, we argue that the pose information was not well utilized and hasn't yet been fully exploited for person ReID. In this work, we introduce a novel framework called Attention-Aware Compositional Network (AACN) for person ReID. AACN consists of two main components: Pose-guided Part Attention (PPA) and Attention-aware Feature Composition (AFC). PPA is learned and applied to mask out undesirable background features in pedestrian feature maps. Furthermore, pose-guided visibility scores are estimated for body parts to deal with part occlusion in the proposed AFC module. Extensive experiments with ablation analysis show the effectiveness of our method, and state-of-the-art results are achieved on several public datasets, including Market-1501, CUHK03, CUHK01, SenseReID, CUHK03-NP and DukeMTMC-reID.

This paper describes our proposed method for the Content-Based Video Relevance Prediction (CBVRP) challenge. Our method is based on deep learning, i.e. we train a deep network to predict the relevance between two video sequences from their features. We explore the usage of second-order relevance, both in preparing training data, and in extending the deep network. Second-order relevance refers to e.g. the relevance between x and z if x is relevant to y and y is relevant to z. In our proposed method, we use second-order relevance to increase positive samples and decrease negative samples, when preparing training data. We further extend the deep network with an attention module, where the attention mechanism is designed for second-order relevant video sequences. We verify the effectiveness of our method on the validation set of the CBVRP challenge.

Human eyes can recognize person identities based on small salient regions, i.e., person saliency is distinctive and reliable in pedestrian matching across disjoint camera views. However, such valuable information is often hidden when computing similarities of pedestrian images with existing approaches. Inspired by our user study result of human perception on person saliency, we propose a novel perspective for person re-identification based on learning person saliency and matching saliency distribution. The proposed saliency learning and matching framework consists of four steps: (1) To handle misalignment caused by drastic viewpoint change and pose variations, we apply adjacency constrained patch matching to build dense correspondence between image pairs. (2) We propose two alternative methods, i.e., K-Nearest Neighbors and One-class SVM, to estimate a saliency score for each image patch, through which distinctive features stand out without using identity labels in the training procedure. (3) saliency matching is proposed based on patch matching. Matching patches with inconsistent saliency brings penalty, and images of the same identity are recognized by minimizing the saliency matching cost. (4) Furthermore, saliency matching is tightly integrated with patch matching in a unified structural RankSVM learning framework. The effectiveness of our approach is validated on the four public datasets. Our approach outperforms the state-of-the-art person re-identification methods on all these datasets.

Low-level saliency cues or priors do not produce good enough saliency detection results especially when the salient object presents in a low-contrast background with confusing visual appearance. This issue raises a serious problem for conventional approaches. In this paper, we tackle this problem by proposing a multi-context deep learning framework for salient object detection. We employ deep Convolutional Neural Networks to model saliency of objects in images. Global context and local context are both taken into account, and are jointly modeled in a unified multi-context deep learning framework. To provide a better initialization for training the deep neural networks, we investigate different pre-training strategies, and a task-specific pre-training scheme is designed to make the multi-context modeling suited for saliency detection. Furthermore, recently proposed contemporary deep models in the ImageNet Image Classification Challenge are tested, and their effectiveness in saliency detection are investigated. Our approach is extensively evaluated on five public datasets, and experimental results show significant and consistent improvements over the state-of-the-art methods.

Person re-identification is to match pedestrian images from disjoint camera views detected by pedestrian detectors. Challenges are presented in the form of complex variations of lightings, poses, viewpoints, blurring effects, image resolutions, camera settings, occlusions and background clutter across camera views. In addition, misalignment introduced by the pedestrian detector will affect most existing person re-identification methods that use manually cropped pedestrian images and assume perfect detection. In this paper, we propose a novel filter pairing neural network (FPNN) to jointly handle misalignment, photometric and geometric transforms, occlusions and background clutter. All the key components are jointly optimized to maximize the strength of each component when cooperating with others. In contrast to existing works that use handcrafted features, our method automatically learns features optimal for the re-identification task from data. The learned filter pairs encode photometric transforms. Its deep architecture makes it possible to model a mixture of complex photometric and geometric transforms. We build the largest benchmark re-id dataset with 13, 164 images of 1, 360 pedestrians. Unlike existing datasets, which only provide manually cropped pedestrian images, our dataset provides automatically detected bounding boxes for evaluation close to practical applications. Our neural network significantly outperforms state-of-the-art methods on this dataset.

In this paper, we propose a novel approach of learning mid-level filters from automatically discovered patch clusters for person re-identification. It is well motivated by our study on what are good filters for person re-identification. Our mid-level filters are discriminatively learned for identifying specific visual patterns and distinguishing persons, and have good cross-view invariance. First, local patches are qualitatively measured and classified with their discriminative power. Discriminative and representative patches are collected for filter learning. Second, patch clusters with coherent appearance are obtained by pruning hierarchical clustering trees, and a simple but effective cross-view training strategy is proposed to learn filters that are view-invariant and discriminative. Third, filter responses are integrated with patch matching scores in RankSVM training. The effectiveness of our approach is validated on the VIPeR dataset and the CUHK01 dataset. The learned mid-level features are complementary to existing handcrafted low-level features, and improve the best Rank-1 matching rate on the VIPeR dataset by 14%.

Human eyes can recognize person identities based on some small salient regions. However, such valuable salient information is often hidden when computing similarities of images with existing approaches. Moreover, many existing approaches learn discriminative features and handle drastic viewpoint change in a supervised way and require labeling new training data for a different pair of camera views. In this paper, we propose a novel perspective for person re-identification based on unsupervised salience learning. Distinctive features are extracted without requiring identity labels in the training procedure. First, we apply adjacency constrained patch matching to build dense correspondence between image pairs, which shows effectiveness in handling misalignment caused by large viewpoint and pose variations. Second, we learn human salience in an unsupervised manner. To improve the performance of person re-identification, human salience is incorporated in patch matching to find reliable and discriminative matched patches. The effectiveness of our approach is validated on the widely used VIPeR dataset and ETHZ dataset.

Human salience is distinctive and reliable information in matching pedestrians across disjoint camera views. In this paper, we exploit the pair wise salience distribution relationship between pedestrian images, and solve the person re-identification problem by proposing a salience matching strategy. To handle the misalignment problem in pedestrian images, patch matching is adopted and patch salience is estimated. Matching patches with inconsistent salience brings penalty. Images of the same person are recognized by minimizing the salience matching cost. Furthermore, our salience matching is tightly integrated with patch matching in a unified structural Rank SVM learning framework. The effectiveness of our approach is validated on the VIPeR dataset and the CUHK Campus dataset. It outperforms the state-of-the-art methods on both datasets.

Automatically counting vehicles in complex traffic scenes from videos is challenging. Detection and tracking algorithms may fail due to occlusions, scene clutters, and large variations of viewpoints and vehicle types. We propose a new approach of counting vehicles through exploiting contextual regularities from scene structures. It breaks the problem into simpler problems, which count vehicles on each path separately. The model of each path and its source and sink add strong regularization on the motion and the sizes of vehicles and can thus significantly improve the accuracy of vehicle counting. Our approach is based on tracking and clustering feature points and can be summarized in threefold. First, an algorithm is proposed to automatically learn the models of scene structures. A traffic scene is segmented into local semantic regions by exploiting the temporal cooccurrence of local motions. Local semantic regions are connected into global complete paths using the proposed fast marching algorithm. Sources and sinks are estimated from the models of semantic regions. Second, an algorithm is proposed to cluster trajectories of feature points into objects and to estimate average vehicle sizes at different locations from initial clustering results. Third, trajectories of features points are often fragmented due to occlusions. By integrating the spatiotemporal features of trajectory clusters with contextual models of paths and sources and sinks, trajectory clusters are assigned into different paths and connected into complete trajectories. Experimental results on a complex traffic scene show the effectiveness of our approach.

Efficient linear filtering in DCT domain is important in the area of processing and manipulation of image and video streams compressed in DCT-based method. In this paper, we proposed a novel method for linear filtering in DCT domain, regardless of filter type. We decompose any filter by SVD into weighted separable sub-filters which are well studied. Then we do fast linear filtering using these separable subfilters in DCT domain, and combine their results. To our best knowledge, it is the first method capable to do linear filtering with any type of filters directly in DCT domain. The scheme is demonstrated and discussed by doing Gabor filtering in DCT domain. Experiment results show that convolution result using the proposed solution is the same as that in spatial domain. Furthermore, our scheme is well suitable for distributed computing, which will improve computing speed greatly.