Wei Ke

Yanhao Wu, Tong Zhang, Wei Ke et al.

Self-supervised learning (SSL) has the potential to benefit many applications, particularly those where manually annotating data is cumbersome. One such situation is the semantic segmentation of point clouds. In this context, existing methods employ contrastive learning strategies and define positive pairs by performing various augmentation of point clusters in a single frame. As such, these methods do not exploit the temporal nature of LiDAR data. In this paper, we introduce an SSL strategy that leverages positive pairs in both the spatial and temporal domain. To this end, we design (i) a point-to-cluster learning strategy that aggregates spatial information to distinguish objects; and (ii) a cluster-to-cluster learning strategy based on unsupervised object tracking that exploits temporal correspondences. We demonstrate the benefits of our approach via extensive experiments performed by self-supervised training on two large-scale LiDAR datasets and transferring the resulting models to other point cloud segmentation benchmarks. Our results evidence that our method outperforms the state-of-the-art point cloud SSL methods.

Zicheng Zhang, Yi Zhu, Jianzhuang Liu et al.

Referring image segmentation aims at localizing all pixels of the visual objects described by a natural language sentence. Previous works learn to straightforwardly align the sentence embedding and pixel-level embedding for highlighting the referred objects, but ignore the semantic consistency of pixels within the same object, leading to incomplete masks and localization errors in predictions. To tackle this problem, we propose CoupAlign, a simple yet effective multi-level visual-semantic alignment method, to couple sentence-mask alignment with word-pixel alignment to enforce object mask constraint for achieving more accurate localization and segmentation. Specifically, the Word-Pixel Alignment (WPA) module performs early fusion of linguistic and pixel-level features in intermediate layers of the vision and language encoders. Based on the word-pixel aligned embedding, a set of mask proposals are generated to hypothesize possible objects. Then in the Sentence-Mask Alignment (SMA) module, the masks are weighted by the sentence embedding to localize the referred object, and finally projected back to aggregate the pixels for the target. To further enhance the learning of the two alignment modules, an auxiliary loss is designed to contrast the foreground and background pixels. By hierarchically aligning pixels and masks with linguistic features, our CoupAlign captures the pixel coherence at both visual and semantic levels, thus generating more accurate predictions. Extensive experiments on popular datasets (e.g., RefCOCO and G-Ref) show that our method achieves consistent improvements over state-of-the-art methods, e.g., about 2% oIoU increase on the validation and testing set of RefCOCO. Especially, CoupAlign has remarkable ability in distinguishing the target from multiple objects of the same class.

Congpei Qiu, Tong Zhang, Wei Ke et al.

Dense Self-Supervised Learning (SSL) methods address the limitations of using image-level feature representations when handling images with multiple objects. Although the dense features extracted by employing segmentation maps and bounding boxes allow networks to perform SSL for each object, we show that they suffer from coupling and positional bias, which arise from the receptive field increasing with layer depth and zero-padding. We address this by introducing three data augmentation strategies, and leveraging them in (i) a decoupling module that aims to robustify the network to variations in the object's surroundings, and (ii) a de-positioning module that encourages the network to discard positional object information. We demonstrate the benefits of our method on COCO and on a new challenging benchmark, OpenImage-MINI, for object classification, semantic segmentation, and object detection. Our extensive experiments evidence the better generalization of our method compared to the SOTA dense SSL methods

Xiangyu Xiong, Yue Sun, Xiaohong Liu et al.

Although current data augmentation methods are successful to alleviate the data insufficiency, conventional augmentation are primarily intra-domain while advanced generative adversarial networks (GANs) generate images remaining uncertain, particularly in small-scale datasets. In this paper, we propose a parameterized GAN (ParaGAN) that effectively controls the changes of synthetic samples among domains and highlights the attention regions for downstream classification. Specifically, ParaGAN incorporates projection distance parameters in cyclic projection and projects the source images to the decision boundary to obtain the class-difference maps. Our experiments show that ParaGAN can consistently outperform the existing augmentation methods with explainable classification on two small-scale medical datasets.

Yao Xiao, Weiyan Chen, Jiahao Chen et al.

Current AI-Generated Image (AIGI) detection approaches predominantly rely on binary classification to distinguish real from synthetic images, often lacking interpretable or convincing evidence to substantiate their decisions. This limitation stems from existing AIGI detection benchmarks, which, despite featuring a broad collection of synthetic images, remain restricted in their coverage of artifact diversity and lack detailed, localized annotations. To bridge this gap, we introduce a fine-grained benchmark towards eXplainable AI-Generated image Detection, named X-AIGD, which provides pixel-level, categorized annotations of perceptual artifacts, spanning low-level distortions, high-level semantics, and cognitive-level counterfactuals. These comprehensive annotations facilitate fine-grained interpretability evaluation and deeper insight into model decision-making processes. Our extensive investigation using X-AIGD provides several key insights: (1) Existing AIGI detectors demonstrate negligible reliance on perceptual artifacts, even at the most basic distortion level. (2) While AIGI detectors can be trained to identify specific artifacts, they still substantially base their judgment on uninterpretable features. (3) Explicitly aligning model attention with artifact regions can increase the interpretability and generalization of detectors. The data and code are available at: https://github.com/Coxy7/X-AIGD.

Jie Hou, Jiakang Li, Xiaokang Peng et al.

Community detection becomes an important problem with the booming of social networks. The Medoid-Shift algorithm preserves the benefits of Mean-Shift and can be applied to problems based on distance matrix, such as community detection. One drawback of the Medoid-Shift algorithm is that there may be no data points within the neighborhood region defined by a distance parameter. To deal with the community detection problem better, a new algorithm called Revised Medoid-Shift (RMS) in this work is thus proposed. During the process of finding the next medoid, the RMS algorithm is based on a neighborhood defined by KNN, while the original Medoid-Shift is based on a neighborhood defined by a distance parameter. Since the neighborhood defined by KNN is more stable than the one defined by the distance parameter in terms of the number of data points within the neighborhood, the RMS algorithm may converge more smoothly. In the RMS method, each of the data points is shifted towards a medoid within the neighborhood defined by KNN. After the iterative process of shifting, each of the data point converges into a cluster center, and the data points converging into the same center are grouped into the same cluster. The RMS algorithm is tested on two kinds of datasets including community datasets with known ground truth partition and community datasets without ground truth partition respectively. The experiment results show sthat the proposed RMS algorithm generally produces betster results than Medoid-Shift and some state-of-the-art together with most classic community detection algorithms on different kinds of community detection datasets.

Congpei Qiu, Zhaoyu Hu, Wei Ke et al.

Representation learning with Vision Transformers (ViTs) has advanced rapidly, yet the utility of large-scale models in spatially sensitive tasks is hindered by spurious tokens. Prior efforts to mitigate this have been limited, often defining these artifacts narrowly, for example, as simple high-norm outliers. We argue that this scope is insufficient. For dense prediction tasks, we posit that any token failing to encode location-aligned semantics should be treated as a spurious artifact. This broader definition reveals a more complex problem, leading us to systematically categorize and characterize three fundamental types of spurious tokens that corrupt spatial representations. Based on this comprehensive diagnosis, we propose UniRefiner, a universal refinement framework that teaches pre-trained ViTs to self-dispose of these artifacts. UniRefiner uses contrastive registers to explicitly isolate and redistribute spurious tokens via a dual objective: (i) it aligns image tokens with filtered regular tokens to preserve semantics, and (ii) it aligns register tokens with detected spurious tokens to capture the spurious signals. Our method requires only a few epochs of fine-tuning on ~5k images to refine diverse ViTs, including massive models like EVA-CLIP-8B and InternViT-6B. Experiments demonstrate consistent and significant improvements: notably, the refined EVA-CLIP-8B achieves 51.9\% mIoU on ADE20K (+9.4\%), surpassing specialized vision models like DINOv2 (49.1\%), while zero-shot segmentation accuracy improves by up to 22\%. UniRefiner unlocks the latent spatial potential of existing large-scale foundation models, paving the way for their broader application.

Shiwei Zhang, Zhengzheng Wang, Qing Liu et al.

This paper presents a new annotation method called Sparse Annotation (SA) for crowd counting, which reduces human labeling efforts by sparsely labeling individuals in an image. We argue that sparse labeling can reduce the redundancy of full annotation and capture more diverse information from distant individuals that is not fully captured by Partial Annotation methods. Besides, we propose a point-based Progressive Point Matching network (PPM) to better explore the crowd from the whole image with sparse annotation, which includes a Proposal Matching Network (PMN) and a Performance Restoration Network (PRN). The PMN generates pseudo-point samples using a basic point classifier, while the PRN refines the point classifier with the pseudo points to maximize performance. Our experimental results show that PPM outperforms previous semi-supervised crowd counting methods with the same amount of annotation by a large margin and achieves competitive performance with state-of-the-art fully-supervised methods.

Lingzhi Pan, Tong Zhang, Bingyuan Chen et al.

With the advancements in denoising diffusion probabilistic models (DDPMs), image inpainting has significantly evolved from merely filling information based on nearby regions to generating content conditioned on various prompts such as text, exemplar images, and sketches. However, existing methods, such as model fine-tuning and simple concatenation of latent vectors, often result in generation failures due to overfitting and inconsistency between the inpainted region and the background. In this paper, we argue that the current large diffusion models are sufficiently powerful to generate realistic images without further tuning. Hence, we introduce PILOT (in\textbf{P}ainting v\textbf{I}a \textbf{L}atent \textbf{O}p\textbf{T}imization), an optimization approach grounded on a novel \textit{semantic centralization} and \textit{background preservation loss}. Our method searches latent spaces capable of generating inpainted regions that exhibit high fidelity to user-provided prompts while maintaining coherence with the background. Furthermore, we propose a strategy to balance optimization expense and image quality, significantly enhancing generation efficiency. Our method seamlessly integrates with any pre-trained model, including ControlNet and DreamBooth, making it suitable for deployment in multi-modal editing tools. Our qualitative and quantitative evaluations demonstrate that PILOT outperforms existing approaches by generating more coherent, diverse, and faithful inpainted regions in response to provided prompts.

Haiwan Wei, Yitian Yuan, Xiaohan Lan et al.

Despite progress in video large language models (Video-LLMs), research on instructional video understanding, crucial for enhancing access to instructional content, remains insufficient. To address this, we introduce InstructionBench, an Instructional video understanding Benchmark, which challenges models' advanced temporal reasoning within instructional videos characterized by their strict step-by-step flow. Employing GPT-4, we formulate Q&A pairs in open-ended and multiple-choice formats to assess both Coarse-Grained event-level and Fine-Grained object-level reasoning. Our filtering strategies exclude questions answerable purely by common-sense knowledge, focusing on visual perception and analysis when evaluating Video-LLM models. The benchmark finally contains 5k questions across over 700 videos. We evaluate the latest Video-LLMs on our InstructionBench, finding that closed-source models outperform open-source ones. However, even the best model, GPT-4o, achieves only 53.42% accuracy, indicating significant gaps in temporal reasoning. To advance the field, we also develop a comprehensive instructional video dataset with over 19k Q&A pairs from nearly 2.5k videos, using an automated data generation framework, thereby enriching the community's research resources. All data are available at https://huggingface.co/datasets/sunwhw/InstructionBench.

Chak Fong Chong, Xinyi Fang, Jielong Guo et al.

Large-scale image datasets are often partially labeled, where only a few categories' labels are known for each image. Assigning pseudo-labels to unknown labels to gain additional training signals has become prevalent for training deep classification models. However, some pseudo-labels are inevitably incorrect, leading to a notable decline in the model classification performance. In this paper, we propose a novel method called Category-wise Fine-Tuning (CFT), aiming to reduce model inaccuracies caused by the wrong pseudo-labels. In particular, CFT employs known labels without pseudo-labels to fine-tune the logistic regressions of trained models individually to calibrate each category's model predictions. Genetic Algorithm, seldom used for training deep models, is also utilized in CFT to maximize the classification performance directly. CFT is applied to well-trained models, unlike most existing methods that train models from scratch. Hence, CFT is general and compatible with models trained with different methods and schemes, as demonstrated through extensive experiments. CFT requires only a few seconds for each category for calibration with consumer-grade GPUs. We achieve state-of-the-art results on three benchmarking datasets, including the CheXpert chest X-ray competition dataset (ensemble mAUC 93.33%, single model 91.82%), partially labeled MS-COCO (average mAP 83.69%), and Open Image V3 (mAP 85.31%), outperforming the previous bests by 0.28%, 2.21%, 2.50%, and 0.91%, respectively. The single model on CheXpert has been officially evaluated by the competition server, endorsing the correctness of the result. The outstanding results and generalizability indicate that CFT could be substantial and prevalent for classification model development. Code is available at: https://github.com/maxium0526/category-wise-fine-tuning.

Huoren Yang, Jianchao Zhao, Hu Yusong et al.

Vision-Language-Action (VLA) models have advanced rapidly with stronger backbones, broader pre-training, and larger demonstration datasets, yet their action heads remain largely homogeneous: most directly predict action commands in a fixed world coordinate frame. We propose \textbf{MCF-Proto}, a lightweight action head that equips VLA policies with a Motion-Centric Action Frame (MCF) and a prototype-based action parameterization. At each step, the policy predicts a rotation $R_t \in SO(3)$, composes actions in the transformed local frame from a set of prototypes, and maps them back to the world frame for end-to-end training, using only standard demonstrations without auxiliary supervision. This simple design induces stable emergent structure. Without explicit directional labels, the learned local frames develop a stable geometric structure whose axes are strongly compatible with demonstrated end-effector motion. Meanwhile, actions in the learned representation become substantially more compact, with variation captured by fewer dominant directions and more regularly organized by shared prototypes. These structural properties translate into improved robustness, especially under geometric perturbations. Our results suggest that adding lightweight geometric and compositional structure to the action head can materially improve how VLA policies organize and generalize robotic manipulation behavior. An anonymized code repository is provided in the supplementary material.

Qing Liu, Haotian Liu, Wei Ke et al.

Feature reassembly is an essential component in modern CNN-based segmentation approaches, which includes feature downsampling and upsampling operators. Existing operators reassemble multiple features from a small predefined region into one for each target location independently. This may result in loss of spatial information, which could vanish activations caused by tiny lesions particularly when they cluster together. In this paper, we propose a many-to-many reassembly of features (M2MRF). It reassembles features in a dimension-reduced feature space and simultaneously aggregates multiple features inside a large predefined region into multiple target features. In this way, long range spatial dependencies are captured to maintain activations on tiny lesions. Experimental results on two lesion segmentation benchmarks, i.e. DDR and IDRiD, show that (1) our M2MRF outperforms existing feature reassembly operators; (2) equipped with our M2MRF, the HRNetv2 is able to achieve significant better performance to CNN-based segmentation methods and competitive even better performance to two recent transformer-based segmentation methods. Our code is made publicly available at https://github.com/CVIU-CSU/M2MRF-Lesion-Segmentation.

Wei Ke, Jie Chen, Jianbin Jiao et al.

In this paper, we establish a baseline for object symmetry detection in complex backgrounds by presenting a new benchmark and an end-to-end deep learning approach, opening up a promising direction for symmetry detection in the wild. The new benchmark, named Sym-PASCAL, spans challenges including object diversity, multi-objects, part-invisibility, and various complex backgrounds that are far beyond those in existing datasets. The proposed symmetry detection approach, named Side-output Residual Network (SRN), leverages output Residual Units (RUs) to fit the errors between the object symmetry groundtruth and the outputs of RUs. By stacking RUs in a deep-to-shallow manner, SRN exploits the 'flow' of errors among multiple scales to ease the problems of fitting complex outputs with limited layers, suppressing the complex backgrounds, and effectively matching object symmetry of different scales. Experimental results validate both the benchmark and its challenging aspects related to realworld images, and the state-of-the-art performance of our symmetry detection approach. The benchmark and the code for SRN are publicly available at https://github.com/KevinKecc/SRN.

Bowen Liu, Haijian Lai, Chan-Tong Lam et al.

Although many complex models were proposed to analyze time series data, some studies have demonstrated remarkable performance with simpler structures. A recent study proposed a non-parametric framework for 3D point cloud classification, which has the potential to be adapted for time series forecasting and enable interpretability. Inspired by the previous works, we present TSNN, a non-parametric and interpretable framework for traffic time series forecasting. TSNN consists of multiple layers that decouple the time series by matching the entries in a memory bank, where the memory bank is constructed using a similar matching process within the training set. It leverages the periodicity in traffic data to enhance forecasting accuracy while maintaining a simple model architecture. The proposed model operates without trainable parameters, preserving its inherent interpretability. In the experiments, TSNN achieves competitive performance compared to the typical deep learning models in four real-world traffic flow datasets. We also visualize the decoupling process to show the effectiveness of the components. Finally, we demonstrate the interpretability of the model and illustrate the contribution of each time step within the memory bank.

Yanhao Wu, Haoyang Zhang, Fei He et al.

End-to-end autonomous driving has rapidly progressed, enabling joint perception and planning in complex environments. In the planning stage, state-of-the-art (SOTA) end-to-end autonomous driving models decouple planning into parallel lateral and longitudinal predictions. While effective, this parallel design can lead to i) coordination failures between the planned path and speed, and ii) underutilization of the drive path as a prior for longitudinal planning, thus redundantly encoding static information. To address this, we propose a novel cascaded framework that explicitly conditions longitudinal planning on the drive path, enabling coordinated and collision-aware lateral and longitudinal planning. Specifically, we introduce a path-conditioned formulation that explicitly incorporates the drive path into longitudinal planning. Building on this, the model predicts longitudinal displacements along the drive path rather than full 2D trajectory waypoints. This design simplifies longitudinal reasoning and more tightly couples it with lateral planning. Additionally, we introduce a planning-oriented data augmentation strategy that simulates rare safety-critical events, such as vehicle cut-ins, by adding agents and relabeling longitudinal targets to avoid collision. Evaluated on the challenging Bench2Drive benchmark, our method sets a new SOTA, achieving a driving score of 89.07 and a success rate of 73.18%, demonstrating significantly improved coordination and safety

Zicheng Zhang, Tong Zhang, Yi Zhu et al.

The pre-trained vision-language model, exemplified by CLIP, advances zero-shot semantic segmentation by aligning visual features with class embeddings through a transformer decoder to generate semantic masks. Despite its effectiveness, prevailing methods within this paradigm encounter challenges, including overfitting on seen classes and small fragmentation in masks. To mitigate these issues, we propose a Language-Driven Visual Consensus (LDVC) approach, fostering improved alignment of semantic and visual information.Specifically, we leverage class embeddings as anchors due to their discrete and abstract nature, steering vision features toward class embeddings. Moreover, to circumvent noisy alignments from the vision part due to its redundant nature, we introduce route attention into self-attention for finding visual consensus, thereby enhancing semantic consistency within the same object. Equipped with a vision-language prompting strategy, our approach significantly boosts the generalization capacity of segmentation models for unseen classes. Experimental results underscore the effectiveness of our approach, showcasing mIoU gains of 4.5 on the PASCAL VOC 2012 and 3.6 on the COCO-Stuff 164k for unseen classes compared with the state-of-the-art methods.

Jianlei Chang, Ruofeng Mei, Wei Ke et al.

Generative modeling has recently shown remarkable promise for visuomotor policy learning, enabling flexible and expressive control across diverse embodied AI tasks. However, existing generative policies often struggle with data inefficiency, requiring large-scale demonstrations, and sampling inefficiency, incurring slow action generation during inference. We introduce EfficientFlow, a unified framework for efficient embodied AI with flow-based policy learning. To enhance data efficiency, we bring equivariance into flow matching. We theoretically prove that when using an isotropic Gaussian prior and an equivariant velocity prediction network, the resulting action distribution remains equivariant, leading to improved generalization and substantially reduced data demands. To accelerate sampling, we propose a novel acceleration regularization strategy. As direct computation of acceleration is intractable for marginal flow trajectories, we derive a novel surrogate loss that enables stable and scalable training using only conditional trajectories. Across a wide range of robotic manipulation benchmarks, the proposed algorithm achieves competitive or superior performance under limited data while offering dramatically faster inference. These results highlight EfficientFlow as a powerful and efficient paradigm for high-performance embodied AI.

Yanhao Wu, Tong Zhang, Wei Ke et al.

In the realm of point cloud scene understanding, particularly in indoor scenes, objects are arranged following human habits, resulting in objects of certain semantics being closely positioned and displaying notable inter-object correlations. This can create a tendency for neural networks to exploit these strong dependencies, bypassing the individual object patterns. To address this challenge, we introduce a novel self-supervised learning (SSL) strategy. Our approach leverages both object patterns and contextual cues to produce robust features. It begins with the formulation of an object-exchanging strategy, where pairs of objects with comparable sizes are exchanged across different scenes, effectively disentangling the strong contextual dependencies. Subsequently, we introduce a context-aware feature learning strategy, which encodes object patterns without relying on their specific context by aggregating object features across various scenes. Our extensive experiments demonstrate the superiority of our method over existing SSL techniques, further showing its better robustness to environmental changes. Moreover, we showcase the applicability of our approach by transferring pre-trained models to diverse point cloud datasets.

Xiaoying Zhang, Da Peng, Yipeng Zhang et al.

Recent progress in (multimodal) large language models ((M)LLMs) has shifted focus from pre-training to inference-time computation and post-training optimization, largely due to concerns over the availability of high-quality human data. However, these strategies alone are insufficient to drive substantial model improvements. We argue that effective model advancement requires strong synergy among pre-training, inference-time computation, and post-training optimization. In this paper, we introduce Self-Improving cognition (SIcog), a self-learning framework for constructing next-generation foundation MLLMs by imparting multimodal knowledge and enhancing systematic cognitive capabilities through multimodal pre-training with self-generated data. Specifically, we propose Chain-of-Description for step-by-step visual understanding and integrate structured Chain-of-Thought (CoT) reasoning to support in-depth multimodal reasoning. SIcog first equips a base model with systematic perception and reasoning using minimal external supervision. The enhanced models then generate candidate image captions and CoT reasoning responses for unlabeled images and image-question pairs across diverse tasks, which are filtered through a semantic-similarity-guided self-consistency mechanism. These high-quality, self-generated samples enable large-scale multimodal pre-training, creating a self-improvement loop. Experiments demonstrate SIcog's effectiveness in developing MLLMs with enhanced multimodal cognition. Using only 213K self-generated pre-training samples, SIcog achieves significant improvements, including +3.6% on MMStar and +3.5% on AI2D, outperforming previous pre-training approaches. When combined with post-training techniques for CoT reasoning, SIcog yields +9% gains on MMVet and +8.5% on ScienceQA.

Congpei Qiu, Yanhao Wu, Wei Ke et al.

Contrastive Language-Image Pre-training (CLIP) excels in global alignment with language but exhibits limited sensitivity to spatial information, leading to strong performance in zero-shot classification tasks but underperformance in tasks requiring precise spatial understanding. Recent approaches have introduced Region-Language Alignment (RLA) to enhance CLIP's performance in dense multimodal tasks by aligning regional visual representations with corresponding text inputs. However, we find that CLIP ViTs fine-tuned with RLA suffer from notable loss in spatial awareness, which is crucial for dense prediction tasks. To address this, we propose the Spatial Correlation Distillation (SCD) framework, which preserves CLIP's inherent spatial structure and mitigates the above degradation. To further enhance spatial correlations, we introduce a lightweight Refiner that extracts refined correlations directly from CLIP before feeding them into SCD, based on an intriguing finding that CLIP naturally captures high-quality dense features. Together, these components form a robust distillation framework that enables CLIP ViTs to integrate both visual-language and visual-centric improvements, achieving state-of-the-art results across various open-vocabulary dense prediction benchmarks.

Yanhao Wu, Haoyang Zhang, Tianwei Lin et al.

Generative models in Autonomous Driving (AD) enable diverse scene creation, yet existing methods fall short by only capturing a limited range of modalities, restricting the capability of generating controllable scenes for comprehensive evaluation of AD systems. In this paper, we introduce a multimodal generation framework that incorporates four major data modalities, including a novel addition of map modality. With tokenized modalities, our scene sequence generation framework autoregressively predicts each scene while managing computational demands through a two-stage approach. The Temporal AutoRegressive (TAR) component captures inter-frame dynamics for each modality while the Ordered AutoRegressive (OAR) component aligns modalities within each scene by sequentially predicting tokens in a fixed order. To maintain coherence between map and ego-action modalities, we introduce the Action-aware Map Alignment (AMA) module, which applies a transformation based on the ego-action to maintain coherence between these modalities. Our framework effectively generates complex, realistic driving scenes over extended sequences, ensuring multimodal consistency and offering fine-grained control over scene elements. Project page: https://yanhaowu.github.io/UMGen/

Yichen Zhang, Da Peng, Zonghao Guo et al.

A recent cutting-edge topic in multimodal modeling is to unify visual comprehension and generation within a single model. However, the two tasks demand mismatched decoding regimes and visual representations, making it non-trivial to jointly optimize within a shared feature space. In this work, we present Cheers, a unified multimodal model that decouples patch-level details from semantic representations, thereby stabilizing semantics for multimodal understanding and improving fidelity for image generation via gated detail residuals. Cheers includes three key components: (i) a unified vision tokenizer that encodes and compresses image latent states into semantic tokens for efficient LLM conditioning, (ii) an LLM-based Transformer that unifies autoregressive decoding for text generation and diffusion decoding for image generation, and (iii) a cascaded flow matching head that decodes visual semantics first and then injects semantically gated detail residuals from the vision tokenizer to refine high-frequency content. Experiments on popular benchmarks demonstrate that Cheers matches or surpasses advanced UMMs in both visual understanding and generation. Cheers also achieves 4x token compression, enabling more efficient high-resolution image encoding and generation. Notably, Cheers outperforms the Tar-1.5B on the popular benchmarks GenEval and MMBench, while requiring only 20% of the training cost, indicating effective and efficient (i.e., 4x token compression) unified multimodal modeling. We will release all code and data for future research.

Zewei Wu, César Teixeira, Wei Ke et al.

Visual pedestrian tracking represents a promising research field, with extensive applications in intelligent surveillance, behavior analysis, and human-computer interaction. However, real-world applications face significant occlusion challenges. When multiple pedestrians interact or overlap, the loss of target features severely compromises the tracker's ability to maintain stable trajectories. Traditional tracking methods, which typically rely on full-body bounding box features extracted from {Re-ID} models and linear constant-velocity motion assumptions, often struggle in severe occlusion scenarios. To address these limitations, this work proposes an enhanced tracking framework that leverages richer feature representations and a more robust motion model. Specifically, the proposed method incorporates detection features from both the regression and classification branches of an object detector, embedding spatial and positional information directly into the feature representations. To further mitigate occlusion challenges, a head keypoint detection model is introduced, as the head is less prone to occlusion compared to the full body. In terms of motion modeling, we propose an iterative Kalman filtering approach designed to align with modern detector assumptions, integrating 3D priors to better complete motion trajectories in complex scenes. By combining these advancements in appearance and motion modeling, the proposed method offers a more robust solution for multi-object tracking in crowded environments where occlusions are prevalent.

Zewei Wu, Longhao Wang, Cui Wang et al.

Tracking specific targets, such as pedestrians and vehicles, has been the focus of recent vision-based multitarget tracking studies. However, in some real-world scenarios, unseen categories often challenge existing methods due to low-confidence detections, weak motion and appearance constraints, and long-term occlusions. To address these issues, this article proposes a tracklet-enhanced tracker called Multi-Tracklet Tracking (MTT) that integrates flexible tracklet generation into a multi-tracklet association framework. This framework first adaptively clusters the detection results according to their short-term spatio-temporal correlation into robust tracklets and then estimates the best tracklet partitions using multiple clues, such as location and appearance over time to mitigate error propagation in long-term association. Finally, extensive experiments on the benchmark for generic multiple object tracking demonstrate the competitiveness of the proposed framework.

Jiale Wang, Chen Zhao, Wei Ke et al.

Random Sample Consensus (RANSAC) is a fundamental approach for robustly estimating parametric models from noisy data. Existing learning-based RANSAC methods utilize deep learning to enhance the robustness of RANSAC against outliers. However, these approaches are trained and tested on the data generated by the same algorithms, leading to limited generalization to out-of-distribution data during inference. Therefore, in this paper, we introduce a novel diffusion-based paradigm that progressively injects noise into ground-truth data, simulating the noisy conditions for training learning-based RANSAC. To enhance data diversity, we incorporate Monte Carlo sampling into the diffusion paradigm, approximating diverse data distributions by introducing different types of randomness at multiple stages. We evaluate our approach in the context of feature matching through comprehensive experiments on the ScanNet and MegaDepth datasets. The experimental results demonstrate that our Monte Carlo diffusion mechanism significantly improves the generalization ability of learning-based RANSAC. We also develop extensive ablation studies that highlight the effectiveness of key components in our framework.

Chak Fong Chong, Jielong Guo, Xu Yang et al.

Multi-label image classification datasets are often partially labeled where many labels are missing, posing a significant challenge to training accurate deep classifiers. However, the powerful Mixup sample-mixing data augmentation cannot be well utilized to address this challenge, as it cannot perform linear interpolation on the unknown labels to construct augmented samples. In this paper, we propose LogicMix, a Mixup variant designed for such partially labeled datasets. LogicMix mixes the sample labels by logical OR so that the unknown labels can be correctly mixed by utilizing OR's logical equivalences, including the domination and identity laws. Unlike Mixup, which mixes exactly two samples, LogicMix can mix multiple ($\geq2$) partially labeled samples, constructing visually more confused augmented samples to regularize training. LogicMix is more general and effective than other compared Mixup variants in the experiments on various partially labeled dataset scenarios. Moreover, it is plug-and-play and only requires minimal computation, hence it can be easily inserted into existing frameworks to collaborate with other methods to improve model performance with a negligible impact on training time, as demonstrated through extensive experiments. In particular, through the collaboration of LogicMix, RandAugment, Curriculum Labeling, and Category-wise Fine-Tuning, we attain state-of-the-art performance on MS-COCO, VG-200, and Pascal VOC 2007 benchmarking datasets. The remarkable generality, effectiveness, collaboration, and simplicity suggest that LogicMix promises to be a popular and vital data augmentation method.

Xiangyu Xiong, Yue Sun, Xiaohong Liu et al.

Despite the potential benefits of data augmentation for mitigating the data insufficiency, traditional augmentation methods primarily rely on the prior intra-domain knowledge. On the other hand, advanced generative adversarial networks (GANs) generate inter-domain samples with limited variety. These previous methods make limited contributions to describing the decision boundaries for binary classification. In this paper, we propose a distance guided GAN (DisGAN) which controls the variation degrees of generated samples in the hyperplane space. Specifically, we instantiate the idea of DisGAN by combining two ways. The first way is vertical distance GAN (VerDisGAN) where the inter-domain generation is conditioned on the vertical distances. The second way is horizontal distance GAN (HorDisGAN) where the intra-domain generation is conditioned on the horizontal distances. Furthermore, VerDisGAN can produce the class-specific regions by mapping the source images to the hyperplane. Experimental results show that DisGAN consistently outperforms the GAN-based augmentation methods with explainable binary classification. The proposed method can apply to different classification architectures and has potential to extend to multi-class classification.

Tong Zhang, Congpei Qiu, Wei Ke et al.

Self-supervised learning (SSL) methods aim to learn view-invariant representations by maximizing the similarity between the features extracted from different crops of the same image regardless of cropping size and content. In essence, this strategy ignores the fact that two crops may truly contain different image information, e.g., background and small objects, and thus tends to restrain the diversity of the learned representations. In this work, we address this issue by introducing a new self-supervised learning strategy, LoGo, that explicitly reasons about Local and Global crops. To achieve view invariance, LoGo encourages similarity between global crops from the same image, as well as between a global and a local crop. However, to correctly encode the fact that the content of smaller crops may differ entirely, LoGo promotes two local crops to have dissimilar representations, while being close to global crops. Our LoGo strategy can easily be applied to existing SSL methods. Our extensive experiments on a variety of datasets and using different self-supervised learning frameworks validate its superiority over existing approaches. Noticeably, we achieve better results than supervised models on transfer learning when using only 1/10 of the data.

Shiwei Zhang, Wei Ke, Lin Yang

Weakly supervised object detection (WSOD) is a challenging task that requires simultaneously learn object classifiers and estimate object locations under the supervision of image category labels. A major line of WSOD methods roots in multiple instance learning which regards images as bags of instances and selects positive instances from each bag to learn the detector. However, a grand challenge emerges when the detector inclines to converge to discriminative parts of objects rather than the whole objects. In this paper, under the hypothesis that optimal solutions are included in local minima, we propose a discovery-and-selection approach fused with multiple instance learning (DS-MIL), which finds rich local minima and select optimal solution from multiple local minima. To implement DS-MIL, an attention module is proposed so that more context information can be captured by feature maps and more valuable proposals can be collected during training. With proposal candidates, a selection module is proposed to select informative instances for object detector. Experimental results on commonly used benchmarks show that our proposed DS-MIL approach can consistently improve the baselines, reporting state-of-the-art performance.

Ka-Hou Chan, Sio-Kei Im, Wei Ke

This article introduces a multiple classifier method to improve the performance of concatenate-designed neural networks, such as ResNet and DenseNet, with the purpose to alleviate the pressure on the final classifier. We give the design of the classifiers, which collects the features produced between the network sets, and present the constituent layers and the activation function for the classifiers, to calculate the classification score of each classifier. We use the L2 normalization method to obtain the classifier score instead of the Softmax normalization. We also determine the conditions that can enhance convergence. As a result, the proposed classifiers are able to improve the accuracy in the experimental cases significantly, and show that the method not only has better performance than the original models, but also produces faster convergence. Moreover, our classifiers are general and can be applied to all classification related concatenate-designed network models.

Zhekun Luo, Devin Guillory, Baifeng Shi et al.

Weakly-supervised action localization requires training a model to localize the action segments in the video given only video level action label. It can be solved under the Multiple Instance Learning (MIL) framework, where a bag (video) contains multiple instances (action segments). Since only the bag's label is known, the main challenge is assigning which key instances within the bag to trigger the bag's label. Most previous models use attention-based approaches applying attentions to generate the bag's representation from instances, and then train it via the bag's classification. These models, however, implicitly violate the MIL assumption that instances in negative bags should be uniformly negative. In this work, we explicitly model the key instances assignment as a hidden variable and adopt an Expectation-Maximization (EM) framework. We derive two pseudo-label generation schemes to model the E and M process and iteratively optimize the likelihood lower bound. We show that our EM-MIL approach more accurately models both the learning objective and the MIL assumptions. It achieves state-of-the-art performance on two standard benchmarks, THUMOS14 and ActivityNet1.2.

Wei Ke, Tianliang Zhang, Zeyi Huang et al.

Classification and localization are two pillars of visual object detectors. However, in CNN-based detectors, these two modules are usually optimized under a fixed set of candidate (or anchor) bounding boxes. This configuration significantly limits the possibility to jointly optimize classification and localization. In this paper, we propose a Multiple Instance Learning (MIL) approach that selects anchors and jointly optimizes the two modules of a CNN-based object detector. Our approach, referred to as Multiple Anchor Learning (MAL), constructs anchor bags and selects the most representative anchors from each bag. Such an iterative selection process is potentially NP-hard to optimize. To address this issue, we solve MAL by repetitively depressing the confidence of selected anchors by perturbing their corresponding features. In an adversarial selection-depression manner, MAL not only pursues optimal solutions but also fully leverages multiple anchors/features to learn a detection model. Experiments show that MAL improves the baseline RetinaNet with significant margins on the commonly used MS-COCO object detection benchmark and achieves new state-of-the-art detection performance compared with recent methods.

Qing Liu, Xiaopeng Hong, Wei Ke et al.

Existing joint optic disc and cup segmentation approaches are developed either in Cartesian or polar coordinate system. However, due to the subtle optic cup, the contextual information exploited from the single domain even by the prevailing CNNs is still insufficient. In this paper, we propose a novel segmentation approach, named Cartesian-polar dual-domain network (DDNet), which for the first time considers the complementary of the Cartesian domain and the polar domain. We propose a two-branch of domain feature encoder and learn translation equivariant representations on rectilinear grid from Cartesian domain and rotation equivariant representations on polar grid from polar domain parallelly. To fuse the features on two different grids, we propose a dual-domain fusion module. This module builds the correspondence between two grids by the differentiable polar transform layer and learns the feature importance across two domains in element-wise to enhance the expressive capability. Finally, the decoder aggregates the fused features from low-level to high-level and makes dense predictions. We validate the state-of-the-art segmentation performances of our DDNet on the public dataset ORIGA. According to the segmentation masks, we estimate the commonly used clinical measure for glaucoma, i.e., the vertical cup-to-disc ratio. The low cup-to-disc ratio estimation error demonstrates the potential application in glaucoma screening.

Fang Wan, Chang Liu, Wei Ke et al.

Weakly supervised object detection (WSOD) is a challenging task when provided with image category supervision but required to simultaneously learn object locations and object detectors. Many WSOD approaches adopt multiple instance learning (MIL) and have non-convex loss functions which are prone to get stuck into local minima (falsely localize object parts) while missing full object extent during training. In this paper, we introduce a continuation optimization method into MIL and thereby creating continuation multiple instance learning (C-MIL), with the intention of alleviating the non-convexity problem in a systematic way. We partition instances into spatially related and class related subsets, and approximate the original loss function with a series of smoothed loss functions defined within the subsets. Optimizing smoothed loss functions prevents the training procedure falling prematurely into local minima and facilitates the discovery of Stable Semantic Extremal Regions (SSERs) which indicate full object extent. On the PASCAL VOC 2007 and 2012 datasets, C-MIL improves the state-of-the-art of weakly supervised object detection and weakly supervised object localization with large margins.

Zeyi Huang, Wei Ke, Dong Huang

Improving object detectors against occlusion, blur and noise is a critical step to deploy detectors in real applications. Since it is not possible to exhaust all image defects through data collection, many researchers seek to generate hard samples in training. The generated hard samples are either images or feature maps with coarse patches dropped out in the spatial dimensions. Significant overheads are required in training the extra hard samples and/or estimating drop-out patches using extra network branches. In this paper, we improve object detectors using a highly efficient and fine-grain mechanism called Inverted Attention (IA). Different from the original detector network that only focuses on the dominant part of objects, the detector network with IA iteratively inverts attention on feature maps and puts more attention on complementary object parts, feature channels and even context. Our approach (1) operates along both the spatial and channels dimensions of the feature maps; (2) requires no extra training on hard samples, no extra network parameters for attention estimation, and no testing overheads. Experiments show that our approach consistently improved both two-stage and single-stage detectors on benchmark databases.

Yilong Yang, Xiaoshan Li, Zhiming Liu et al.

Prototyping is an effective and efficient way of requirement validation to avoid introducing errors in the early stage of software development. However, manually developing a prototype of a software system requires additional efforts, which would increase the overall cost of software development. In this paper, we present an approach with a developed tool to automatic generation of prototypes from formal requirements models. A requirements model consists of a use case diagram, a conceptual class diagram, use case definitions specified by system sequence diagrams and the contracts of their system operations. We propose a method to decompose a contract into executable parts and non-executable parts. A set of transformation rules is given to decompose the executable part into pre-implemented primitive operations. A non-executable part is usually realized by significant algorithms such as sorting a list, finding the shortest path or domain-specific computation. It can be implemented manually or by using existing code. A CASE tool is developed that provides an interface for developers to develop a program for each non-executable part of a contract, and automatically transforms the executables into sequences of pre-implemented primitive operations. We have conducted four cases studies with over 50 use cases. The experimental result shows that the 93.65% of requirement specifications are executable, and only 6.35% are non-executable such as sorting and event-call, which can be implemented by developers manually or invoking the APIs of advanced algorithms in Java library. The one second generated the prototype of a case study requires approximate nine hours manual implementation by a skilled programmer. Overall, the result is satisfiable, and the proposed approach with the developed CASE tool can be applied to the software industry for requirements engineering.

Chang Liu, Wei Ke, Fei Qin et al.

Robust object skeleton detection requires to explore rich representative visual features and effective feature fusion strategies. In this paper, we first re-visit the implementation of HED, the essential principle of which can be ideally described with a linear reconstruction model. Hinted by this, we formalize a Linear Span framework, and propose Linear Span Network (LSN) modified by Linear Span Units (LSUs), which minimize the reconstruction error of convolutional network. LSN further utilizes subspace linear span beside the feature linear span to increase the independence of convolutional features and the efficiency of feature integration, which enlarges the capability of fitting complex ground-truth. As a result, LSN can effectively suppress the cluttered backgrounds and reconstruct object skeletons. Experimental results validate the state-of-the-art performance of the proposed LSN.

Wei Ke, Jie Chen, Jianbin Jiao et al.

In this paper, we establish a baseline for object reflection symmetry detection in complex backgrounds by presenting a new benchmark and an end-to-end deep learning approach, opening up a promising direction for symmetry detection in the wild. The new benchmark, Sym-PASCAL, spans challenges including object diversity, multi-objects, part-invisibility, and various complex backgrounds that are far beyond those in existing datasets. The end-to-end deep learning approach, referred to as a side-output residual network (SRN), leverages the output residual units (RUs) to fit the errors between the object ground-truth symmetry and the side-outputs of multiple stages. By cascading RUs in a deep-to-shallow manner, SRN exploits the 'flow' of errors among multiple stages to address the challenges of fitting complex output with limited convolutional layers, suppressing the complex backgrounds, and effectively matching object symmetry at different scales. SRN is further upgraded to a multi-task side-output residual network (MT-SRN) for joint symmetry and edge detection, demonstrating its generality to image-to-mask learning tasks. Experimental results validate both the challenging aspects of Sym-PASCAL benchmark related to real-world images and the state-of-the-art performance of the proposed SRN approach.

Yilong Yang, Xiaoshan Li, Nafees Qamar et al.

Legacy Electronic Health Records (EHRs) systems were not developed with the level of connectivity expected from them nowadays. Therefore, interoperability weakness inherent in the legacy systems can result in poor patient care and waste of financial resources. Large hospitals are less likely to share their data with external hospitals due to economic and political reasons. Motivated by these facts, we aim to provide a set of software implementation guidelines, i.e., MedShare to deal with interoperability issues among disconnected healthcare systems. The proposed integrated architecture includes: 1) a data extractor to fetch legacy medical data from a hemodialysis center, 2) converting it to a common data model, 3) indexing patient information using the HashMap technique, and 4) a set of services and tools that can be installed as a coherent environment on top of stand-alone EHRs systems. Our work enabled three cooperating but autonomous hospitals to mutually exchange medical data and helped them develop a common reference architecture. It lets stakeholders retain control over their patient data, winning the trust and confidence much needed towards a successful deployment of MedShare. Security concerns were effectively addressed that also included patient consent in the data exchange process. Thereby, the implemented toolset offered a collaborative environment to share EHRs by the healthcare providers.