Fang Zhao

Tianfei Zhou, Meijie Zhang, Fang Zhao et al.

Learning semantic segmentation from weakly-labeled (e.g., image tags only) data is challenging since it is hard to infer dense object regions from sparse semantic tags. Despite being broadly studied, most current efforts directly learn from limited semantic annotations carried by individual image or image pairs, and struggle to obtain integral localization maps. Our work alleviates this from a novel perspective, by exploring rich semantic contexts synergistically among abundant weakly-labeled training data for network learning and inference. In particular, we propose regional semantic contrast and aggregation (RCA) . RCA is equipped with a regional memory bank to store massive, diverse object patterns appearing in training data, which acts as strong support for exploration of dataset-level semantic structure. Particularly, we propose i) semantic contrast to drive network learning by contrasting massive categorical object regions, leading to a more holistic object pattern understanding, and ii) semantic aggregation to gather diverse relational contexts in the memory to enrich semantic representations. In this manner, RCA earns a strong capability of fine-grained semantic understanding, and eventually establishes new state-of-the-art results on two popular benchmarks, i.e., PASCAL VOC 2012 and COCO 2014.

Lingdong Kong, Yaru Niu, Shaoyuan Xie et al.

Accurate depth estimation under out-of-distribution (OoD) scenarios, such as adverse weather conditions, sensor failure, and noise contamination, is desirable for safety-critical applications. Existing depth estimation systems, however, suffer inevitably from real-world corruptions and perturbations and are struggled to provide reliable depth predictions under such cases. In this paper, we summarize the winning solutions from the RoboDepth Challenge -- an academic competition designed to facilitate and advance robust OoD depth estimation. This challenge was developed based on the newly established KITTI-C and NYUDepth2-C benchmarks. We hosted two stand-alone tracks, with an emphasis on robust self-supervised and robust fully-supervised depth estimation, respectively. Out of more than two hundred participants, nine unique and top-performing solutions have appeared, with novel designs ranging from the following aspects: spatial- and frequency-domain augmentations, masked image modeling, image restoration and super-resolution, adversarial training, diffusion-based noise suppression, vision-language pre-training, learned model ensembling, and hierarchical feature enhancement. Extensive experimental analyses along with insightful observations are drawn to better understand the rationale behind each design. We hope this challenge could lay a solid foundation for future research on robust and reliable depth estimation and beyond. The datasets, competition toolkit, workshop recordings, and source code from the winning teams are publicly available on the challenge website.

Wenhao Wang, Fang Zhao, Shengcai Liao et al.

Unsupervised domain adaptation (UDA) for person re-identification is challenging because of the huge gap between the source and target domain. A typical self-training method is to use pseudo-labels generated by clustering algorithms to iteratively optimize the model on the target domain. However, a drawback to this is that noisy pseudo-labels generally cause trouble in learning. To address this problem, a mutual learning method by dual networks has been developed to produce reliable soft labels. However, as the two neural networks gradually converge, their complementarity is weakened and they likely become biased towards the same kind of noise. This paper proposes a novel light-weight module, the Attentive WaveBlock (AWB), which can be integrated into the dual networks of mutual learning to enhance the complementarity and further depress noise in the pseudo-labels. Specifically, we first introduce a parameter-free module, the WaveBlock, which creates a difference between features learned by two networks by waving blocks of feature maps differently. Then, an attention mechanism is leveraged to enlarge the difference created and discover more complementary features. Furthermore, two kinds of combination strategies, i.e. pre-attention and post-attention, are explored. Experiments demonstrate that the proposed method achieves state-of-the-art performance with significant improvements on multiple UDA person re-identification tasks. We also prove the generality of the proposed method by applying it to vehicle re-identification and image classification tasks. Our codes and models are available at https://github.com/WangWenhao0716/Attentive-WaveBlock.

Weijia Li, Guanglei Chu, Jiong Chen et al.

Recent advances in industrial anomaly detection have highlighted the need for deeper logical anomaly analysis, where unexpected relationships among objects, counts, and spatial configurations must be identified and explained. Existing approaches often rely on large-scale external reasoning modules or elaborate pipeline designs, hindering practical deployment and interpretability. To address these limitations, we introduce a new task, Reasoning Logical Anomaly Detection (RLAD), which extends traditional anomaly detection by incorporating logical reasoning. We propose a new framework, LAD-Reasoner, a customized tiny multimodal language model built on Qwen2.5-VL 3B. Our approach leverages a two-stage training paradigm that first employs Supervised Fine-Tuning (SFT) for fine-grained visual understanding, followed by Group Relative Policy Optimization (GRPO) to refine logical anomaly detection and enforce coherent, human-readable reasoning. Crucially, reward signals are derived from both the detection accuracy and the structural quality of the outputs, obviating the need for building chain of thought (CoT) reasoning data. Experiments on the MVTec LOCO AD dataset show that LAD-Reasoner, though significantly smaller, matches the performance of Qwen2.5-VL-72B in accuracy and F1 score, and further excels in producing concise and interpretable rationales. This unified design reduces reliance on large models and complex pipelines, while offering transparent and interpretable insights into logical anomaly detection. Code and data will be released.

Hongsong Wang, Wanjiang Weng, Junbo Wang et al.

Human action understanding serves as a foundational pillar in the field of intelligent motion perception. Skeletons serve as a modality- and device-agnostic representation for human modeling, and skeleton-based action understanding has potential applications in humanoid robot control and interaction. \RED{However, existing works often lack the scalability and generalization required to handle diverse action understanding tasks. There is no skeleton foundation model that can be adapted to a wide range of action understanding tasks}. This paper presents a Unified Skeleton-based Dense Representation Learning (USDRL) framework, which serves as a foundational model for skeleton-based human action understanding. USDRL consists of a Transformer-based Dense Spatio-Temporal Encoder (DSTE), Multi-Grained Feature Decorrelation (MG-FD), and Multi-Perspective Consistency Training (MPCT). The DSTE module adopts two parallel streams to learn temporal dynamic and spatial structure features. The MG-FD module collaboratively performs feature decorrelation across temporal, spatial, and instance domains to reduce dimensional redundancy and enhance information extraction. The MPCT module employs both multi-view and multi-modal self-supervised consistency training. The former enhances the learning of high-level semantics and mitigates the impact of low-level discrepancies, while the latter effectively facilitates the learning of informative multimodal features. We perform extensive experiments on 25 benchmarks across across 9 skeleton-based action understanding tasks, covering coarse prediction, dense prediction, and transferred prediction. Our approach significantly outperforms the current state-of-the-art methods. We hope that this work would broaden the scope of research in skeleton-based action understanding and encourage more attention to dense prediction tasks.

Runzhi Deng, Yundi Hu, Xinshuang Zhang et al.

Few-shot multi-class industrial anomaly detection remains a challenging task. Vision-language models need to be both category-adaptive and sharply discriminative, yet data scarcity often blurs the boundary between normal and abnormal states. This ambiguity leads to missed subtle defects and the rejection of atypical normal samples. We propose ABounD, an Adversarial Boundary-Driven few-shot learning for multi-class anomaly detection, which is a unified learning framework that integrates semantic concept learning with decision boundary shaping. The Dynamic Concept Fusion (DCF) module produces class-adaptive prompts by fusing generalizable priors with class-specific cues, conditioned on image features. Meanwhile, Adversarial Boundary Forging (ABF) sculpts a more precise decision margin by generating boundary-level fence features via PGD-style perturbations. Training is conducted in a single stage under a Concept-Boundary Loss, where ABF provides the main supervisory signal and semantic-spatial regularizers stabilize the optimization. This synergy yields a decision boundary that closely follows normal data while preserving flexibility and robust semantic alignment. Experiments on MVTec-AD and VisA datasets demonstrate state-of-the-art performance in the task of few-shot multi-class anomaly detection.

Yuchen Fang, Yanjun Qin, Haiyong Luo et al.

Traffic forecasting is important in intelligent transportation systems of webs and beneficial to traffic safety, yet is very challenging because of the complex and dynamic spatio-temporal dependencies in real-world traffic systems. Prior methods use the pre-defined or learnable static graph to extract spatial correlations. However, the static graph-based methods fail to mine the evolution of the traffic network. Researchers subsequently generate the dynamic graph for each time slice to reflect the changes of spatial correlations, but they follow the paradigm of independently modeling spatio-temporal dependencies, ignoring the cross-time spatial influence. In this paper, we propose a novel cross-time dynamic graph-based deep learning model, named CDGNet, for traffic forecasting. The model is able to effectively capture the cross-time spatial dependence between each time slice and its historical time slices by utilizing the cross-time dynamic graph. Meanwhile, we design a gating mechanism to sparse the cross-time dynamic graph, which conforms to the sparse spatial correlations in the real world. Besides, we propose a novel encoder-decoder architecture to incorporate the cross-time dynamic graph-based GCN for multi-step traffic forecasting. Experimental results on three real-world public traffic datasets demonstrate that CDGNet outperforms the state-of-the-art baselines. We additionally provide a qualitative study to analyze the effectiveness of our architecture.

Yanjun Qin, Yuchen Fang, Haiyong Luo et al.

Traffic forecasting is a problem of intelligent transportation systems (ITS) and crucial for individuals and public agencies. Therefore, researches pay great attention to deal with the complex spatio-temporal dependencies of traffic system for accurate forecasting. However, there are two challenges: 1) Most traffic forecasting studies mainly focus on modeling correlations of neighboring sensors and ignore correlations of remote sensors, e.g., business districts with similar spatio-temporal patterns; 2) Prior methods which use static adjacency matrix in graph convolutional networks (GCNs) are not enough to reflect the dynamic spatial dependence in traffic system. Moreover, fine-grained methods which use self-attention to model dynamic correlations of all sensors ignore hierarchical information in road networks and have quadratic computational complexity. In this paper, we propose a novel dynamic multi-graph convolution recurrent network (DMGCRN) to tackle above issues, which can model the spatial correlations of distance, the spatial correlations of structure, and the temporal correlations simultaneously. We not only use the distance-based graph to capture spatial information from nodes are close in distance but also construct a novel latent graph which encoded the structure correlations among roads to capture spatial information from nodes are similar in structure. Furthermore, we divide the neighbors of each sensor into coarse-grained regions, and dynamically assign different weights to each region at different times. Meanwhile, we integrate the dynamic multi-graph convolution network into the gated recurrent unit (GRU) to capture temporal dependence. Extensive experiments on three real-world traffic datasets demonstrate that our proposed algorithm outperforms state-of-the-art baselines.

Yuchen Fang, Yanjun Qin, Haiyong Luo et al.

Traffic prediction has gradually attracted the attention of researchers because of the increase in traffic big data. Therefore, how to mine the complex spatio-temporal correlations in traffic data to predict traffic conditions more accurately become a difficult problem. Previous works combined graph convolution networks (GCNs) and self-attention mechanism with deep time series models (e.g. recurrent neural networks) to capture the spatio-temporal correlations separately, ignoring the relationships across time and space. Besides, GCNs are limited by over-smoothing issue and self-attention is limited by quadratic problem, result in GCNs lack global representation capabilities, and self-attention inefficiently capture the global spatial dependence. In this paper, we propose a novel deep learning model for traffic forecasting, named Multi-Context Aware Spatio-Temporal Joint Linear Attention (STJLA), which applies linear attention to the spatio-temporal joint graph to capture global dependence between all spatio-temporal nodes efficiently. More specifically, STJLA utilizes static structural context and dynamic semantic context to improve model performance. The static structure context based on node2vec and one-hot encoding enriches the spatio-temporal position information. Furthermore, the multi-head diffusion convolution network based dynamic spatial context enhances the local spatial perception ability, and the GRU based dynamic temporal context stabilizes sequence position information of the linear attention, respectively. Experiments on two real-world traffic datasets, England and PEMSD7, demonstrate that our STJLA can achieve up to 9.83% and 3.08% accuracy improvement in MAE measure over state-of-the-art baselines.

Wenhao Wang, Shengcai Liao, Fang Zhao et al.

Existing person re-identification models often have low generalizability, which is mostly due to limited availability of large-scale labeled data in training. However, labeling large-scale training data is very expensive and time-consuming, while large-scale synthetic dataset shows promising value in learning generalizable person re-identification models. Therefore, in this paper a novel and practical person re-identification task is proposed,i.e. how to use labeled synthetic dataset and unlabeled real-world dataset to train a universal model. In this way, human annotations are no longer required, and it is scalable to large and diverse real-world datasets. To address the task, we introduce a framework with high generalizability, namely DomainMix. Specifically, the proposed method firstly clusters the unlabeled real-world images and selects the reliable clusters. During training, to address the large domain gap between two domains, a domain-invariant feature learning method is proposed, which introduces a new loss,i.e. domain balance loss, to conduct an adversarial learning between domain-invariant feature learning and domain discrimination, and meanwhile learns a discriminative feature for person re-identification. This way, the domain gap between synthetic and real-world data is much reduced, and the learned feature is generalizable thanks to the large-scale and diverse training data. Experimental results show that the proposed annotation-free method is more or less comparable to the counterpart trained with full human annotations, which is quite promising. In addition, it achieves the current state of the art on several person re-identification datasets under direct cross-dataset evaluation.

Jian Zhao, Jianshu Li, Xiaoguang Tu et al.

In this paper, we study the challenging unconstrained set-based face recognition problem where each subject face is instantiated by a set of media (images and videos) instead of a single image. Naively aggregating information from all the media within a set would suffer from the large intra-set variance caused by heterogeneous factors (e.g., varying media modalities, poses and illuminations) and fail to learn discriminative face representations. A novel Multi-Prototype Network (MPNet) model is thus proposed to learn multiple prototype face representations adaptively from the media sets. Each learned prototype is representative for the subject face under certain condition in terms of pose, illumination and media modality. Instead of handcrafting the set partition for prototype learning, MPNet introduces a Dense SubGraph (DSG) learning sub-net that implicitly untangles inconsistent media and learns a number of representative prototypes. Qualitative and quantitative experiments clearly demonstrate superiority of the proposed model over state-of-the-arts.

Jian Zhao, Yu Cheng, Yi Cheng et al.

Despite the remarkable progress in face recognition related technologies, reliably recognizing faces across ages still remains a big challenge. The appearance of a human face changes substantially over time, resulting in significant intra-class variations. As opposed to current techniques for age-invariant face recognition, which either directly extract age-invariant features for recognition, or first synthesize a face that matches target age before feature extraction, we argue that it is more desirable to perform both tasks jointly so that they can leverage each other. To this end, we propose a deep Age-Invariant Model (AIM) for face recognition in the wild with three distinct novelties. First, AIM presents a novel unified deep architecture jointly performing cross-age face synthesis and recognition in a mutual boosting way. Second, AIM achieves continuous face rejuvenation/aging with remarkable photorealistic and identity-preserving properties, avoiding the requirement of paired data and the true age of testing samples. Third, we develop effective and novel training strategies for end-to-end learning the whole deep architecture, which generates powerful age-invariant face representations explicitly disentangled from the age variation. Moreover, we propose a new large-scale Cross-Age Face Recognition (CAFR) benchmark dataset to facilitate existing efforts and push the frontiers of age-invariant face recognition research. Extensive experiments on both our CAFR and several other cross-age datasets (MORPH, CACD and FG-NET) demonstrate the superiority of the proposed AIM model over the state-of-the-arts. Benchmarking our model on one of the most popular unconstrained face recognition datasets IJB-C additionally verifies the promising generalizability of AIM in recognizing faces in the wild.

Jianshu Li, Shengtao Xiao, Fang Zhao et al.

Face analytics benefits many multimedia applications. It consists of a number of tasks, such as facial emotion recognition and face parsing, and most existing approaches generally treat these tasks independently, which limits their deployment in real scenarios. In this paper we propose an integrated Face Analytics Network (iFAN), which is able to perform multiple tasks jointly for face analytics with a novel carefully designed network architecture to fully facilitate the informative interaction among different tasks. The proposed integrated network explicitly models the interactions between tasks so that the correlations between tasks can be fully exploited for performance boost. In addition, to solve the bottleneck of the absence of datasets with comprehensive training data for various tasks, we propose a novel cross-dataset hybrid training strategy. It allows "plug-in and play" of multiple datasets annotated for different tasks without the requirement of a fully labeled common dataset for all the tasks. We experimentally show that the proposed iFAN achieves state-of-the-art performance on multiple face analytics tasks using a single integrated model. Specifically, iFAN achieves an overall F-score of 91.15% on the Helen dataset for face parsing, a normalized mean error of 5.81% on the MTFL dataset for facial landmark localization and an accuracy of 45.73% on the BNU dataset for emotion recognition with a single model.

Fang Zhao, Jiashi Feng, Jian Zhao et al.

Face recognition techniques have been developed significantly in recent years. However, recognizing faces with partial occlusion is still challenging for existing face recognizers which is heavily desired in real-world applications concerning surveillance and security. Although much research effort has been devoted to developing face de-occlusion methods, most of them can only work well under constrained conditions, such as all the faces are from a pre-defined closed set. In this paper, we propose a robust LSTM-Autoencoders (RLA) model to effectively restore partially occluded faces even in the wild. The RLA model consists of two LSTM components, which aims at occlusion-robust face encoding and recurrent occlusion removal respectively. The first one, named multi-scale spatial LSTM encoder, reads facial patches of various scales sequentially to output a latent representation, and occlusion-robustness is achieved owing to the fact that the influence of occlusion is only upon some of the patches. Receiving the representation learned by the encoder, the LSTM decoder with a dual channel architecture reconstructs the overall face and detects occlusion simultaneously, and by feat of LSTM, the decoder breaks down the task of face de-occlusion into restoring the occluded part step by step. Moreover, to minimize identify information loss and guarantee face recognition accuracy over recovered faces, we introduce an identity-preserving adversarial training scheme to further improve RLA. Extensive experiments on both synthetic and real datasets of faces with occlusion clearly demonstrate the effectiveness of our proposed RLA in removing different types of facial occlusion at various locations. The proposed method also provides significantly larger performance gain than other de-occlusion methods in promoting recognition performance over partially-occluded faces.