Changwei Wang

Rongtao Xu, Changwei Wang, Jiaxi Sun et al.

Efficiently training accurate deep models for weakly supervised semantic segmentation (WSSS) with image-level labels is challenging and important. Recently, end-to-end WSSS methods have become the focus of research due to their high training efficiency. However, current methods suffer from insufficient extraction of comprehensive semantic information, resulting in low-quality pseudo-labels and sub-optimal solutions for end-to-end WSSS. To this end, we propose a simple and novel Self Correspondence Distillation (SCD) method to refine pseudo-labels without introducing external supervision. Our SCD enables the network to utilize feature correspondence derived from itself as a distillation target, which can enhance the network's feature learning process by complementing semantic information. In addition, to further improve the segmentation accuracy, we design a Variation-aware Refine Module to enhance the local consistency of pseudo-labels by computing pixel-level variation. Finally, we present an efficient end-to-end Transformer-based framework (TSCD) via SCD and Variation-aware Refine Module for the accurate WSSS task. Extensive experiments on the PASCAL VOC 2012 and MS COCO 2014 datasets demonstrate that our method significantly outperforms other state-of-the-art methods. Our code is available at {https://github.com/Rongtao-Xu/RepresentationLearning/tree/main/SCD-AAAI2023}.

Rongtao Xu, Changwei Wang, Jiguang Zhang et al.

Accurate skin lesion segmentation from dermoscopic images is of great importance for skin cancer diagnosis. However, automatic segmentation of melanoma remains a challenging task because it is difficult to incorporate useful texture representations into the learning process. Texture representations are not only related to the local structural information learned by CNN, but also include the global statistical texture information of the input image. In this paper, we propose a trans\textbf{Former} network (\textbf{SkinFormer}) that efficiently extracts and fuses statistical texture representation for \textbf{Skin} lesion segmentation. Specifically, to quantify the statistical texture of input features, a Kurtosis-guided Statistical Counting Operator is designed. We propose Statistical Texture Fusion Transformer and Statistical Texture Enhance Transformer with the help of Kurtosis-guided Statistical Counting Operator by utilizing the transformer's global attention mechanism. The former fuses structural texture information and statistical texture information, and the latter enhances the statistical texture of multi-scale features. {Extensive experiments on three publicly available skin lesion datasets validate that our SkinFormer outperforms other SOAT methods, and our method achieves 93.2\% Dice score on ISIC 2018. It can be easy to extend SkinFormer to segment 3D images in the future.} Our code is available at https://github.com/Rongtao-Xu/SkinFormer.

Jinzhou Lin, Han Gao, Xuxiang Feng et al.

In recent years, the rapid advancement of Large Language Models (LLMs) such as the Generative Pre-trained Transformer (GPT) has attracted increasing attention due to their potential in a variety of practical applications. The application of LLMs with Embodied Intelligence has emerged as a significant area of focus. Among the myriad applications of LLMs, navigation tasks are particularly noteworthy because they demand a deep understanding of the environment and quick, accurate decision-making. LLMs can augment embodied intelligence systems with sophisticated environmental perception and decision-making support, leveraging their robust language and image-processing capabilities. This article offers an exhaustive summary of the symbiosis between LLMs and embodied intelligence with a focus on navigation. It reviews state-of-the-art models, research methodologies, and assesses the advantages and disadvantages of existing embodied navigation models and datasets. Finally, the article elucidates the role of LLMs in embodied intelligence, based on current research, and forecasts future directions in the field. A comprehensive list of studies in this survey is available at https://github.com/Rongtao-Xu/Awesome-LLM-EN.

Jingqian Wu, Rongtao Xu, Zach Wood-Doughty et al.

Local feature detection and description play an important role in many computer vision tasks, which are designed to detect and describe keypoints in "any scene" and "any downstream task". Data-driven local feature learning methods need to rely on pixel-level correspondence for training, which is challenging to acquire at scale, thus hindering further improvements in performance. In this paper, we propose SAMFeat to introduce SAM (segment anything model), a fundamental model trained on 11 million images, as a teacher to guide local feature learning and thus inspire higher performance on limited datasets. To do so, first, we construct an auxiliary task of Attention-weighted Semantic Relation Distillation (ASRD), which distillates feature relations with category-agnostic semantic information learned by the SAM encoder into a local feature learning network, to improve local feature description using semantic discrimination. Second, we develop a technique called Weakly Supervised Contrastive Learning Based on Semantic Grouping (WSC), which utilizes semantic groupings derived from SAM as weakly supervised signals, to optimize the metric space of local descriptors. Third, we design an Edge Attention Guidance (EAG) to further improve the accuracy of local feature detection and description by prompting the network to pay more attention to the edge region guided by SAM. SAMFeat's performance on various tasks such as image matching on HPatches, and long-term visual localization on Aachen Day-Night showcases its superiority over previous local features. The release code is available at https://github.com/vignywang/SAMFeat.

Changwei Wang, Rongtao Xu, Yuyang Zhang et al.

Limited by the locality of convolutional neural networks, most existing local features description methods only learn local descriptors with local information and lack awareness of global and surrounding spatial context. In this work, we focus on making local descriptors "look wider to describe better" by learning local Descriptors with More Than just Local information (MTLDesc). Specifically, we resort to context augmentation and spatial attention mechanisms to make our MTLDesc obtain non-local awareness. First, Adaptive Global Context Augmented Module and Diverse Local Context Augmented Module are proposed to construct robust local descriptors with context information from global to local. Second, Consistent Attention Weighted Triplet Loss is designed to integrate spatial attention awareness into both optimization and matching stages of local descriptors learning. Third, Local Features Detection with Feature Pyramid is given to obtain more stable and accurate keypoints localization. With the above innovations, the performance of our MTLDesc significantly surpasses the prior state-of-the-art local descriptors on HPatches, Aachen Day-Night localization and InLoc indoor localization benchmarks.

Zhiwei Xu, Hangyu Mao, Nianmin Zhang et al.

In partially observable multi-agent systems, agents typically only have access to local observations. This severely hinders their ability to make precise decisions, particularly during decentralized execution. To alleviate this problem and inspired by image outpainting, we propose State Inference with Diffusion Models (SIDIFF), which uses diffusion models to reconstruct the original global state based solely on local observations. SIDIFF consists of a state generator and a state extractor, which allow agents to choose suitable actions by considering both the reconstructed global state and local observations. In addition, SIDIFF can be effortlessly incorporated into current multi-agent reinforcement learning algorithms to improve their performance. Finally, we evaluated SIDIFF on different experimental platforms, including Multi-Agent Battle City (MABC), a novel and flexible multi-agent reinforcement learning environment we developed. SIDIFF achieved desirable results and outperformed other popular algorithms.

Shunpeng Chen, Yukun Song, Changwei Wang et al.

Visual Place Recognition (VPR) determines a query image's geographic location by matching it against geotagged databases. However, existing methods struggle with perceptual aliasing caused by irrelevant regions and inefficient re-ranking due to rigid candidate scheduling. To address these issues, we introduce FoL++, a method combining robust discriminative region modeling with adaptive re-ranking. Specifically, we propose a Reliability Estimation Branch to generate spatial reliability maps that explicitly model occlusion resistance. This representation is further optimized by two spatial alignment losses (SAL and SCEL) to effectively align features and highlight salient regions. For weakly supervised learning without manual annotations, a pseudo-correspondence strategy generates dense local feature supervision directly from aggregation clusters. Our Adaptive Candidate Scheduler dynamically resizes candidate pools based on global similarity. By weighting local matches by reliability and adaptively fusing global and local evidence, FoL++ surpasses traditional independent matching systems. Extensive experiments across seven benchmarks demonstrate that FoL++ achieves state-of-the-art performance with a lightweight memory footprint, improving inference speed by 40% over FoL. Code and models will be released (and merged with FoL) at https://github.com/chenshunpeng/FoL.

Wenhao Xu, Rongtao Xu, Changwei Wang et al.

Accurate segmentation of colorectal polyps in colonoscopy images is crucial for effective diagnosis and management of colorectal cancer (CRC). However, current deep learning-based methods primarily rely on fusing RGB information across multiple scales, leading to limitations in accurately identifying polyps due to restricted RGB domain information and challenges in feature misalignment during multi-scale aggregation. To address these limitations, we propose the Polyp Segmentation Network with Shunted Transformer (PSTNet), a novel approach that integrates both RGB and frequency domain cues present in the images. PSTNet comprises three key modules: the Frequency Characterization Attention Module (FCAM) for extracting frequency cues and capturing polyp characteristics, the Feature Supplementary Alignment Module (FSAM) for aligning semantic information and reducing misalignment noise, and the Cross Perception localization Module (CPM) for synergizing frequency cues with high-level semantics to achieve efficient polyp segmentation. Extensive experiments on challenging datasets demonstrate PSTNet's significant improvement in polyp segmentation accuracy across various metrics, consistently outperforming state-of-the-art methods. The integration of frequency domain cues and the novel architectural design of PSTNet contribute to advancing computer-assisted polyp segmentation, facilitating more accurate diagnosis and management of CRC.

Wenhao Xu, Changwei Wang, Xuxiang Feng et al.

Vision-language models (VLMs) have demonstrated remarkable open-vocabulary object recognition capabilities, motivating their adaptation for dense prediction tasks like segmentation. However, directly applying VLMs to such tasks remains challenging due to their lack of pixel-level granularity and the limited data available for fine-tuning, leading to overfitting and poor generalization. To address these limitations, we propose Generalization Boosted Adapter (GBA), a novel adapter strategy that enhances the generalization and robustness of VLMs for open-vocabulary segmentation. GBA comprises two core components: (1) a Style Diversification Adapter (SDA) that decouples features into amplitude and phase components, operating solely on the amplitude to enrich the feature space representation while preserving semantic consistency; and (2) a Correlation Constraint Adapter (CCA) that employs cross-attention to establish tighter semantic associations between text categories and target regions, suppressing irrelevant low-frequency ``noise'' information and avoiding erroneous associations. Through the synergistic effect of the shallow SDA and the deep CCA, GBA effectively alleviates overfitting issues and enhances the semantic relevance of feature representations. As a simple, efficient, and plug-and-play component, GBA can be flexibly integrated into various CLIP-based methods, demonstrating broad applicability and achieving state-of-the-art performance on multiple open-vocabulary segmentation benchmarks.

Zhe Feng, Sen Lian, Changwei Wang et al.

The quadratic complexity of softmax attention presents a major obstacle for scaling Transformers to high-resolution vision tasks. Existing linear attention variants often replace the softmax with Gaussian kernels to reduce complexity, but such approximations lack theoretical grounding and tend to oversuppress mid-range token interactions. We propose LaplacianFormer, a Transformer variant that employs a Laplacian kernel as a principled alternative to softmax, motivated by empirical observations and theoretical analysis. To address expressiveness degradation under low-rank approximations, we introduce a provably injective feature map that retains fine-grained token information. For efficient computation, we adopt a Nyström approximation of the kernel matrix and solve the resulting system using Newton--Schulz iteration, avoiding costly matrix inversion and SVD. We further develop custom CUDA implementations for both the kernel and solver, enabling high-throughput forward and backward passes suitable for edge deployment. Experiments on ImageNet show that LaplacianFormer achieves strong performance-efficiency trade-offs while improving attention expressiveness.

Shibiao Xu, ShuChen Zheng, Wenhao Xu et al.

Infrared small object detection is an important computer vision task involving the recognition and localization of tiny objects in infrared images, which usually contain only a few pixels. However, it encounters difficulties due to the diminutive size of the objects and the generally complex backgrounds in infrared images. In this paper, we propose a deep learning method, HCF-Net, that significantly improves infrared small object detection performance through multiple practical modules. Specifically, it includes the parallelized patch-aware attention (PPA) module, dimension-aware selective integration (DASI) module, and multi-dilated channel refiner (MDCR) module. The PPA module uses a multi-branch feature extraction strategy to capture feature information at different scales and levels. The DASI module enables adaptive channel selection and fusion. The MDCR module captures spatial features of different receptive field ranges through multiple depth-separable convolutional layers. Extensive experimental results on the SIRST infrared single-frame image dataset show that the proposed HCF-Net performs well, surpassing other traditional and deep learning models. Code is available at https://github.com/zhengshuchen/HCFNet.

Shibiao Xu, Shunpeng Chen, Rongtao Xu et al.

Local feature matching enjoys wide-ranging applications in the realm of computer vision, encompassing domains such as image retrieval, 3D reconstruction, and object recognition. However, challenges persist in improving the accuracy and robustness of matching due to factors like viewpoint and lighting variations. In recent years, the introduction of deep learning models has sparked widespread exploration into local feature matching techniques. The objective of this endeavor is to furnish a comprehensive overview of local feature matching methods. These methods are categorized into two key segments based on the presence of detectors. The Detector-based category encompasses models inclusive of Detect-then-Describe, Joint Detection and Description, Describe-then-Detect, as well as Graph Based techniques. In contrast, the Detector-free category comprises CNN Based, Transformer Based, and Patch Based methods. Our study extends beyond methodological analysis, incorporating evaluations of prevalent datasets and metrics to facilitate a quantitative comparison of state-of-the-art techniques. The paper also explores the practical application of local feature matching in diverse domains such as Structure from Motion, Remote Sensing Image Registration, and Medical Image Registration, underscoring its versatility and significance across various fields. Ultimately, we endeavor to outline the current challenges faced in this domain and furnish future research directions, thereby serving as a reference for researchers involved in local feature matching and its interconnected domains. A comprehensive list of studies in this survey is available at https://github.com/vignywang/Awesome-Local-Feature-Matching .

Run Ling, Ke Cao, Jian Lu et al.

Multi-subject video generation aims to synthesize videos from textual prompts and multiple reference images, ensuring that each subject preserves natural scale and visual fidelity. However, current methods face two challenges: scale inconsistency, where variations in subject size lead to unnatural generation, and permutation sensitivity, where the order of reference inputs causes subject distortion. In this paper, we propose MoFu, a unified framework that tackles both challenges. For scale inconsistency, we introduce Scale-Aware Modulation (SMO), an LLM-guided module that extracts implicit scale cues from the prompt and modulates features to ensure consistent subject sizes. To address permutation sensitivity, we present a simple yet effective Fourier Fusion strategy that processes the frequency information of reference features via the Fast Fourier Transform to produce a unified representation. Besides, we design a Scale-Permutation Stability Loss to jointly encourage scale-consistent and permutation-invariant generation. To further evaluate these challenges, we establish a dedicated benchmark with controlled variations in subject scale and reference permutation. Extensive experiments demonstrate that MoFu significantly outperforms existing methods in preserving natural scale, subject fidelity, and overall visual quality.

Xingtian Pei, Yukun Song, Changwei Wang et al.

Zero-Shot Compositional Image Retrieval (ZS-CIR) requires both preserving the visual continuity of the reference image and faithfully executing the semantic variables specified in the modification text, which constitutes the core challenge of the task. Existing methods often suffer from Perception Myopia in a single space, or fall into Logic Drift in iterative collaboration due to the perception ceiling of the underlying retriever. To address this issue, we propose a one-stop hierarchical Perception-to-Deliberation Framework (PDF), which, to the best of our knowledge, is the first to introduce experience self-evolution and Test-Time Scaling Law (TTS) into ZS-CIR. Relying on a hierarchical multi-agent architecture, PDF first utilizes an Intent Routing Manager to dynamically dispatch multi-view Worker perception signals based on modification intents to construct a high-recall candidate pool. Subsequently, the Decision Manager combines a Training-free Reasoning Policy Distillation mechanism with a Tournament-style TTS strategy to achieve self-evolving fine-grained reasoning, yielding the final retrieval results. Experimental results demonstrate that PDF achieves SOTA performance on three benchmark datasets: CIRR, CIRCO, and FashionIQ. This study indicates that experience-driven self-evolution and TTS represent a highly promising and scalable path for achieving zero-shot fine-grained multimedia retrieval. The code will be made publicly available upon acceptance.

Xiaofeng Han, Shunpeng Chen, Zenghuang Fu et al.

Robot vision has greatly benefited from advancements in multimodal fusion techniques and vision-language models (VLMs). We adopt a task-oriented perspective to systematically review the applications and advancements of multimodal fusion methods and VLMs in the field of robot vision. For semantic scene understanding tasks, we categorize fusion approaches into encoder-decoder frameworks, attention-based architectures, and graph neural networks. Meanwhile, we also analyze the architectural characteristics and practical implementations of these fusion strategies in key tasks such as simultaneous localization and mapping (SLAM), 3D object detection, navigation, and manipulation. We compare the evolutionary paths and applicability of VLMs based on large language models (LLMs) with traditional multimodal fusion methods.Additionally, we conduct an in-depth analysis of commonly used datasets, evaluating their applicability and challenges in real-world robotic scenarios. Building on this analysis, we identify key challenges in current research, including cross-modal alignment, efficient fusion, real-time deployment, and domain adaptation. We propose future directions such as self-supervised learning for robust multimodal representations, structured spatial memory and environment modeling to enhance spatial intelligence, and the integration of adversarial robustness and human feedback mechanisms to enable ethically aligned system deployment. Through a comprehensive review, comparative analysis, and forward-looking discussion, we provide a valuable reference for advancing multimodal perception and interaction in robotic vision. A comprehensive list of studies in this survey is available at https://github.com/Xiaofeng-Han-Res/MF-RV.

Zixuan Gong, Guangyin Bao, Qi Zhang et al.

Reconstruction of static visual stimuli from non-invasion brain activity fMRI achieves great success, owning to advanced deep learning models such as CLIP and Stable Diffusion. However, the research on fMRI-to-video reconstruction remains limited since decoding the spatiotemporal perception of continuous visual experiences is formidably challenging. We contend that the key to addressing these challenges lies in accurately decoding both high-level semantics and low-level perception flows, as perceived by the brain in response to video stimuli. To the end, we propose NeuroClips, an innovative framework to decode high-fidelity and smooth video from fMRI. NeuroClips utilizes a semantics reconstructor to reconstruct video keyframes, guiding semantic accuracy and consistency, and employs a perception reconstructor to capture low-level perceptual details, ensuring video smoothness. During inference, it adopts a pre-trained T2V diffusion model injected with both keyframes and low-level perception flows for video reconstruction. Evaluated on a publicly available fMRI-video dataset, NeuroClips achieves smooth high-fidelity video reconstruction of up to 6s at 8FPS, gaining significant improvements over state-of-the-art models in various metrics, e.g., a 128% improvement in SSIM and an 81% improvement in spatiotemporal metrics. Our project is available at https://github.com/gongzix/NeuroClips.

Wenhao Xu, Rongtao Xu, Changwei Wang et al.

Recently, CLIP has found practical utility in the domain of pixel-level zero-shot segmentation tasks. The present landscape features two-stage methodologies beset by issues such as intricate pipelines and elevated computational costs. While current one-stage approaches alleviate these concerns and incorporate Visual Prompt Training (VPT) to uphold CLIP's generalization capacity, they still fall short in fully harnessing CLIP's potential for pixel-level unseen class demarcation and precise pixel predictions. To further stimulate CLIP's zero-shot dense prediction capability, we propose SPT-SEG, a one-stage approach that improves CLIP's adaptability from image to pixel. Specifically, we initially introduce Spectral Prompt Tuning (SPT), incorporating spectral prompts into the CLIP visual encoder's shallow layers to capture structural intricacies of images, thereby enhancing comprehension of unseen classes. Subsequently, we introduce the Spectral Guided Decoder (SGD), utilizing both high and low-frequency information to steer the network's spatial focus towards more prominent classification features, enabling precise pixel-level prediction outcomes. Through extensive experiments on two public datasets, we demonstrate the superiority of our method over state-of-the-art approaches, performing well across all classes and particularly excelling in handling unseen classes. Code is available at:https://github.com/clearxu/SPT.

Changwei Wang, Shunpeng Chen, Yukun Song et al.

Visual Place Recognition (VPR) is aimed at predicting the location of a query image by referencing a database of geotagged images. For VPR task, often fewer discriminative local regions in an image produce important effects while mundane background regions do not contribute or even cause perceptual aliasing because of easy overlap. However, existing methods lack precisely modeling and full exploitation of these discriminative regions. In this paper, we propose the Focus on Local (FoL) approach to stimulate the performance of image retrieval and re-ranking in VPR simultaneously by mining and exploiting reliable discriminative local regions in images and introducing pseudo-correlation supervision. First, we design two losses, Extraction-Aggregation Spatial Alignment Loss (SAL) and Foreground-Background Contrast Enhancement Loss (CEL), to explicitly model reliable discriminative local regions and use them to guide the generation of global representations and efficient re-ranking. Second, we introduce a weakly-supervised local feature training strategy based on pseudo-correspondences obtained from aggregating global features to alleviate the lack of local correspondences ground truth for the VPR task. Third, we suggest an efficient re-ranking pipeline that is efficiently and precisely based on discriminative region guidance. Finally, experimental results show that our FoL achieves the state-of-the-art on multiple VPR benchmarks in both image retrieval and re-ranking stages and also significantly outperforms existing two-stage VPR methods in terms of computational efficiency. Code and models are available at https://github.com/chenshunpeng/FoL

Rongtao Xu, Han Gao, Mingming Yu et al.

With the growing need for diverse and scalable data in indoor scene tasks, such as question answering and dense captioning, we propose 3D-MoRe, a novel paradigm designed to generate large-scale 3D-language datasets by leveraging the strengths of foundational models. The framework integrates key components, including multi-modal embedding, cross-modal interaction, and a language model decoder, to process natural language instructions and 3D scene data. This approach facilitates enhanced reasoning and response generation in complex 3D environments. Using the ScanNet 3D scene dataset, along with text annotations from ScanQA and ScanRefer, 3D-MoRe generates 62,000 question-answer (QA) pairs and 73,000 object descriptions across 1,513 scenes. We also employ various data augmentation techniques and implement semantic filtering to ensure high-quality data. Experiments on ScanQA demonstrate that 3D-MoRe significantly outperforms state-of-the-art baselines, with the CIDEr score improving by 2.15\%. Similarly, on ScanRefer, our approach achieves a notable increase in CIDEr@0.5 by 1.84\%, highlighting its effectiveness in both tasks. Our code and generated datasets will be publicly released to benefit the community, and both can be accessed on the https://3D-MoRe.github.io.

Shide Du, Zihan Fang, Yanchao Tan et al.

Multi-view learning methods leverage multiple data sources to enhance perception by mining correlations across views, typically relying on predefined categories. However, deploying these models in real-world scenarios presents two primary openness challenges. 1) Lack of Interpretability: The integration mechanisms of multi-view data in existing black-box models remain poorly explained; 2) Insufficient Generalization: Most models are not adapted to multi-view scenarios involving unknown categories. To address these challenges, we propose OpenViewer, an openness-aware multi-view learning framework with theoretical support. This framework begins with a Pseudo-Unknown Sample Generation Mechanism to efficiently simulate open multi-view environments and previously adapt to potential unknown samples. Subsequently, we introduce an Expression-Enhanced Deep Unfolding Network to intuitively promote interpretability by systematically constructing functional prior-mapping modules and effectively providing a more transparent integration mechanism for multi-view data. Additionally, we establish a Perception-Augmented Open-Set Training Regime to significantly enhance generalization by precisely boosting confidences for known categories and carefully suppressing inappropriate confidences for unknown ones. Experimental results demonstrate that OpenViewer effectively addresses openness challenges while ensuring recognition performance for both known and unknown samples. The code is released at https://github.com/dushide/OpenViewer.

Wenhao Xu, Shuchen Zheng, Changwei Wang et al.

Infrared small target detection (ISTD) is vital for long-range surveillance in military, maritime, and early warning applications. ISTD is challenged by targets occupying less than 0.15% of the image and low distinguishability from complex backgrounds. Existing deep learning methods often suffer from information loss during downsampling and inefficient global context modeling. This paper presents SAMamba, a novel framework integrating SAM2's hierarchical feature learning with Mamba's selective sequence modeling. Key innovations include: (1) A Feature Selection Adapter (FS-Adapter) for efficient natural-to-infrared domain adaptation via dual-stage selection (token-level with a learnable task embedding and channel-wise adaptive transformations); (2) A Cross-Channel State-Space Interaction (CSI) module for efficient global context modeling with linear complexity using selective state space modeling; and (3) A Detail-Preserving Contextual Fusion (DPCF) module that adaptively combines multi-scale features with a gating mechanism to balance high-resolution and low-resolution feature contributions. SAMamba addresses core ISTD challenges by bridging the domain gap, maintaining fine-grained details, and efficiently modeling long-range dependencies. Experiments on NUAA-SIRST, IRSTD-1k, and NUDT-SIRST datasets show SAMamba significantly outperforms state-of-the-art methods, especially in challenging scenarios with heterogeneous backgrounds and varying target scales. Code: https://github.com/zhengshuchen/SAMamba.

Zherui Zhang, Rongtao Xu, Jie Zhou et al.

The Transformer architecture has achieved significant success in natural language processing, motivating its adaptation to computer vision tasks. Unlike convolutional neural networks, vision transformers inherently capture long-range dependencies and enable parallel processing, yet lack inductive biases and efficiency benefits, facing significant computational and memory challenges that limit its real-world applicability. This paper surveys various online strategies for generating lightweight vision transformers for image recognition, focusing on three key areas: Efficient Component Design, Dynamic Network, and Knowledge Distillation. We evaluate the relevant exploration for each topic on the ImageNet-1K benchmark, analyzing trade-offs among precision, parameters, throughput, and more to highlight their respective advantages, disadvantages, and flexibility. Finally, we propose future research directions and potential challenges in the lightweighting of vision transformers with the aim of inspiring further exploration and providing practical guidance for the community. Project Page: https://github.com/ajxklo/Lightweight-VIT

Shunpeng Chen, Changwei Wang, Rongtao Xu et al.

Visual Place Recognition (VPR) requires robust retrieval of geotagged images despite large appearance, viewpoint, and environmental variation. Prior methods focus on descriptor fine-tuning or fixed sampling strategies yet neglect the dynamic interplay between spatial context and visual similarity during training. We present SAGE (Spatial-visual Adaptive Graph Exploration), a unified training pipeline that enhances granular spatial-visual discrimination by jointly improving local feature aggregation, organize samples during training, and hard sample mining. We introduce a lightweight Soft Probing module that learns residual weights from training data for patch descriptors before bilinear aggregation, boosting distinctive local cues. During training we reconstruct an online geo-visual graph that fuses geographic proximity and current visual similarity so that candidate neighborhoods reflect the evolving embedding landscape. To concentrate learning on the most informative place neighborhoods, we seed clusters from high-affinity anchors and iteratively expand them with a greedy weighted clique expansion sampler. Implemented with a frozen DINOv2 backbone and parameter-efficient fine-tuning, SAGE achieves SOTA across eight benchmarks. It attains 98.9%, 95.8%, 94.5%, and 96.0% Recall@1 on SPED, Pitts30k-test, MSLS-val, and Nordland, respectively. Notably, our method obtains 100% Recall@10 on SPED only using 4096D global descriptors. Code and model will be available at: https://github.com/chenshunpeng/SAGE.

Xiaomeng Zhu, Changwei Wang, Haozhe Wang et al.

A scene graph is a structured represention of objects and their relationships in a scene. Scene Graph Anticipation (SGA) involves predicting future scene graphs from video clips, enabling applications as intelligent surveillance and human-machine collaboration. Existing SGA approaches primarily leverage visual cues, often struggling to integrate valuable commonsense knowledge, thereby limiting long-term prediction robustness. To explicitly leverage such commonsense knowledge, we propose a new approach to better understand the objects, concepts, and relationships in a scene graph. Our approach decouples the SGA task in two steps: first a scene graph capturing model is used to convert a video clip into a sequence of scene graphs, then a pure text-based model is used to predict scene graphs in future frames. Our focus in this work is on the second step, and we call it Linguistic Scene Graph Anticipation (LSGA) and believes it should have independent interest beyond the use in SGA discussed here. For LSGA, we introduce an Object-Oriented Two-Staged Method (OOTSM) where an Large Language Model (LLM) first forecasts object appearances and disappearances before generating detailed human-object relations. We conduct extensive experiments to evaluate OOTSM in two settings. For LSGA, we evaluate our fine-tuned open-sourced LLMs against zero-shot APIs (i.e., GPT-4o, GPT-4o-mini, and DeepSeek-V3) on a benchmark constructed from Action Genome annotations. For SGA, we combine our OOTSM with STTran++ from, and our experiments demonstrate effective state-of-the-art performance: short-term mean-Recall (@10) increases by 3.4% while long-term mean-Recall (@50) improves dramatically by 21.9%. Code is available at https://github.com/ZhuXMMM/OOTSM.

Haoran Xu, Yanlin Liu, Zizhao Tong et al.

Out-of-Distribution (OOD) detection is a critical task that has garnered significant attention. The emergence of CLIP has spurred extensive research into zero-shot OOD detection, often employing a training-free approach. Current methods leverage expert knowledge from large language models (LLMs) to identify potential outliers. However, these approaches tend to over-rely on knowledge in the text space, neglecting the inherent challenges involved in detecting out-of-distribution samples in the image space. In this paper, we propose a novel pipeline, MM-OOD, which leverages the multimodal reasoning capabilities of MLLMs and their ability to conduct multi-round conversations for enhanced outlier detection. Our method is designed to improve performance in both near OOD and far OOD tasks. Specifically, (1) for near OOD tasks, we directly feed ID images and corresponding text prompts into MLLMs to identify potential outliers; and (2) for far OOD tasks, we introduce the sketch-generate-elaborate framework: first, we sketch outlier exposure using text prompts, then generate corresponding visual OOD samples, and finally elaborate by using multimodal prompts. Experiments demonstrate that our method achieves significant improvements on widely used multimodal datasets such as Food-101, while also validating its scalability on ImageNet-1K.

Jiaze Li, Haoran Xu, Wanyi Wu et al.

Federated learning is a paradigm of joint learning in which clients collaborate by sharing model parameters instead of data. However, in the non-iid setting, the global model experiences client drift, which can seriously affect the final performance of the model. Previous methods tend to correct the global model that has already deviated based on the loss function or gradient, overlooking the impact of the client samples. In this paper, we rethink the role of the client side and propose Federated Balanced Learning, i.e., FBL, to prevent this issue from the beginning through sample balance on the client side. Technically, FBL allows unbalanced data on the client side to achieve sample balance through knowledge filling and knowledge sampling using edge-side generation models, under the limitation of a fixed number of data samples on clients. Furthermore, we design a Knowledge Alignment Strategy to bridge the gap between synthetic and real data, and a Knowledge Drop Strategy to regularize our method. Meanwhile, we scale our method to real and complex scenarios, allowing different clients to adopt various methods, and extend our framework to further improve performance. Numerous experiments show that our method outperforms state-of-the-art baselines. The code is released upon acceptance.

Rongtao Xu, Jinzhou Lin, Jialei Zhou et al.

Camera-based occupancy prediction is a mainstream approach for 3D perception in autonomous driving, aiming to infer complete 3D scene geometry and semantics from 2D images. Almost existing methods focus on improving performance through structural modifications, such as lightweight backbones and complex cascaded frameworks, with good yet limited performance. Few studies explore from the perspective of representation fusion, leaving the rich diversity of features in 2D images underutilized. Motivated by this, we propose \textbf{CIGOcc, a two-stage occupancy prediction framework based on multi-level representation fusion. \textbf{CIGOcc extracts segmentation, graphics, and depth features from an input image and introduces a deformable multi-level fusion mechanism to fuse these three multi-level features. Additionally, CIGOcc incorporates knowledge distilled from SAM to further enhance prediction accuracy. Without increasing training costs, CIGOcc achieves state-of-the-art performance on the SemanticKITTI benchmark. The code is provided in the supplementary material and will be released https://github.com/VitaLemonTea1/CIGOcc

Zhiqiang Kou, Junxiang Wu, Wenke Huang et al.

Federated Multi-Label Learning is a distributed paradigm where multiple clients possess heterogeneous multi-label data and perform collaborative learning under privacy constraints without sharing raw data. However, modeling label correlations under heterogeneous distributions remains challenging. Due to client-specific label spaces and varying co-occurrence patterns, correlations learned by individual clients inevitably deviate from the global structure, a phenomenon we term label correlation drift. To address this, we propose FedHarmony, a framework that harmonizes heterogeneous label correlations across clients. It introduces consensus correlation, capturing agreement among other clients and serving as a global teacher to correct biased local estimates. During aggregation, FedHarmony evaluates each client by both data size and correlation quality, assigning weights accordingly. Moreover, we develop an accelerated optimization algorithm for FedHarmony and theoretically establish faster convergence without sacrificing accuracy. Experiments on real-world federated multi-label datasets show that FedHarmony consistently outperforms state-of-the-art methods.

Xiaohong Liu, Xiongkuo Min, Qiang Hu et al.

This paper reports on the NTIRE 2025 XGC Quality Assessment Challenge, which will be held in conjunction with the New Trends in Image Restoration and Enhancement Workshop (NTIRE) at CVPR 2025. This challenge is to address a major challenge in the field of video and talking head processing. The challenge is divided into three tracks, including user generated video, AI generated video and talking head. The user-generated video track uses the FineVD-GC, which contains 6,284 user generated videos. The user-generated video track has a total of 125 registered participants. A total of 242 submissions are received in the development phase, and 136 submissions are received in the test phase. Finally, 5 participating teams submitted their models and fact sheets. The AI generated video track uses the Q-Eval-Video, which contains 34,029 AI-Generated Videos (AIGVs) generated by 11 popular Text-to-Video (T2V) models. A total of 133 participants have registered in this track. A total of 396 submissions are received in the development phase, and 226 submissions are received in the test phase. Finally, 6 participating teams submitted their models and fact sheets. The talking head track uses the THQA-NTIRE, which contains 12,247 2D and 3D talking heads. A total of 89 participants have registered in this track. A total of 225 submissions are received in the development phase, and 118 submissions are received in the test phase. Finally, 8 participating teams submitted their models and fact sheets. Each participating team in every track has proposed a method that outperforms the baseline, which has contributed to the development of fields in three tracks.

Yu Zhang, Jialei Zhou, Xinchen Li et al.

Current text-to-image diffusion generation typically employs complete-text conditioning. Due to the intricate syntax, diffusion transformers (DiTs) inherently suffer from a comprehension defect of complete-text captions. One-fly complete-text input either overlooks critical semantic details or causes semantic confusion by simultaneously modeling diverse semantic primitive types. To mitigate this defect of DiTs, we propose a novel split-text conditioning framework named DiT-ST. This framework converts a complete-text caption into a split-text caption, a collection of simplified sentences, to explicitly express various semantic primitives and their interconnections. The split-text caption is then injected into different denoising stages of DiT-ST in a hierarchical and incremental manner. Specifically, DiT-ST leverages Large Language Models to parse captions, extracting diverse primitives and hierarchically sorting out and constructing these primitives into a split-text input. Moreover, we partition the diffusion denoising process according to its differential sensitivities to diverse semantic primitive types and determine the appropriate timesteps to incrementally inject tokens of diverse semantic primitive types into input tokens via cross-attention. In this way, DiT-ST enhances the representation learning of specific semantic primitive types across different stages. Extensive experiments validate the effectiveness of our proposed DiT-ST in mitigating the complete-text comprehension defect.

Zherui Zhang, Changwei Wang, Rongtao Xu et al.

Data-Free Knowledge Distillation (DFKD) enables the knowledge transfer from the given pre-trained teacher network to the target student model without access to the real training data. Existing DFKD methods focus primarily on improving image recognition performance on associated datasets, often neglecting the crucial aspect of the transferability of learned representations. In this paper, we propose Category-Aware Embedding Data-Free Knowledge Distillation (CAE-DFKD), which addresses at the embedding level the limitations of previous rely on image-level methods to improve model generalization but fail when directly applied to DFKD. The superiority and flexibility of CAE-DFKD are extensively evaluated, including: \textit{\textbf{i.)}} Significant efficiency advantages resulting from altering the generator training paradigm; \textit{\textbf{ii.)}} Competitive performance with existing DFKD state-of-the-art methods on image recognition tasks; \textit{\textbf{iii.)}} Remarkable transferability of data-free learned representations demonstrated in downstream tasks.

Jinzhou Lin, Jie Zhou, Wenhao Xu et al.

Semantic Scene Completion (SSC) aims to infer complete 3D geometry and semantics from monocular images, serving as a crucial capability for camera-based perception in autonomous driving. However, existing SSC methods relying on temporal stacking or depth projection often lack explicit motion reasoning and struggle with occlusions and noisy depth supervision. We propose CurriFlow, a novel semantic occupancy prediction framework that integrates optical flow-based temporal alignment with curriculum-guided depth fusion. CurriFlow employs a multi-level fusion strategy to align segmentation, visual, and depth features across frames using pre-trained optical flow, thereby improving temporal consistency and dynamic object understanding. To enhance geometric robustness, a curriculum learning mechanism progressively transitions from sparse yet accurate LiDAR depth to dense but noisy stereo depth during training, ensuring stable optimization and seamless adaptation to real-world deployment. Furthermore, semantic priors from the Segment Anything Model (SAM) provide category-agnostic supervision, strengthening voxel-level semantic learning and spatial consistency. Experiments on the SemanticKITTI benchmark demonstrate that CurriFlow achieves state-of-the-art performance with a mean IoU of 16.9, validating the effectiveness of our motion-guided and curriculum-aware design for camera-based 3D semantic scene completion.

Jingqian Wu, Peiqi Duan, Zongqiang Wang et al.

In low-light environments, conventional cameras often struggle to capture clear multi-view images of objects due to dynamic range limitations and motion blur caused by long exposure. Event cameras, with their high-dynamic range and high-speed properties, have the potential to mitigate these issues. Additionally, 3D Gaussian Splatting (GS) enables radiance field reconstruction, facilitating bright frame synthesis from multiple viewpoints in low-light conditions. However, naively using an event-assisted 3D GS approach still faced challenges because, in low light, events are noisy, frames lack quality, and the color tone may be inconsistent. To address these issues, we propose Dark-EvGS, the first event-assisted 3D GS framework that enables the reconstruction of bright frames from arbitrary viewpoints along the camera trajectory. Triplet-level supervision is proposed to gain holistic knowledge, granular details, and sharp scene rendering. The color tone matching block is proposed to guarantee the color consistency of the rendered frames. Furthermore, we introduce the first real-captured dataset for the event-guided bright frame synthesis task via 3D GS-based radiance field reconstruction. Experiments demonstrate that our method achieves better results than existing methods, conquering radiance field reconstruction under challenging low-light conditions. The code and sample data are included in the supplementary material.

Yu Zhang, Jingyi Liu, Feng Liu et al.

Visual AutoRegressive modeling (VAR) suffers from substantial computational cost due to the massive token count involved. Failing to account for the continuous evolution of modeling dynamics, existing VAR token reduction methods face three key limitations: heuristic stage partition, non-adaptive schedules, and limited acceleration scope, thereby leaving significant acceleration potential untapped. Since entropy variation intrinsically reflects the transition of predictive uncertainty, it offers a principled measure to capture modeling dynamics evolution. Therefore, we propose NOVA, a training-free token reduction acceleration framework for VAR models via entropy analysis. NOVA adaptively determines the acceleration activation scale during inference by online identifying the inflection point of scale entropy growth. Through scale-linkage and layer-linkage ratio adjustment, NOVA dynamically computes distinct token reduction ratios for each scale and layer, pruning low-entropy tokens while reusing the cache derived from the residuals at the prior scale to accelerate inference and maintain generation quality. Extensive experiments and analyses validate NOVA as a simple yet effective training-free acceleration framework.

Yu Zhang, Jingyi Liu, Yiwei Shi et al.

Visual AutoRegressive modeling (VAR) based on next-scale prediction has revitalized autoregressive visual generation. Although its full-context dependency, i.e., modeling all previous scales for next-scale prediction, facilitates more stable and comprehensive representation learning by leveraging complete information flow, the resulting computational inefficiency and substantial overhead severely hinder VAR's practicality and scalability. This motivates us to develop a new VAR model with better performance and efficiency without full-context dependency. To address this, we reformulate VAR as a non-full-context Markov process, proposing Markov-VAR. It is achieved via Markovian Scale Prediction: we treat each scale as a Markov state and introduce a sliding window that compresses certain previous scales into a compact history vector to compensate for historical information loss owing to non-full-context dependency. Integrating the history vector with the Markov state yields a representative dynamic state that evolves under a Markov process. Extensive experiments demonstrate that Markov-VAR is extremely simple yet highly effective: Compared to VAR on ImageNet, Markov-VAR reduces FID by 10.5% (256 $\times$ 256) and decreases peak memory consumption by 83.8% (1024 $\times$ 1024). We believe that Markov-VAR can serve as a foundation for future research on visual autoregressive generation and other downstream tasks.

Zhiqiang Kou, Junyang Chen, Xin-Qiang Cai et al.

Large language models (LLMs) have achieved impressive results across a range of natural language processing tasks, but their potential to generate harmful content has raised serious safety concerns. Current toxicity detectors primarily rely on single-label benchmarks, which cannot adequately capture the inherently ambiguous and multi-dimensional nature of real-world toxic prompts. This limitation results in biased evaluations, including missed toxic detections and false positives, undermining the reliability of existing detectors. Additionally, gathering comprehensive multi-label annotations across fine-grained toxicity categories is prohibitively costly, further hindering effective evaluation and development. To tackle these issues, we introduce three novel multi-label benchmarks for toxicity detection: \textbf{Q-A-MLL}, \textbf{R-A-MLL}, and \textbf{H-X-MLL}, derived from public toxicity datasets and annotated according to a detailed 15-category taxonomy. We further provide a theoretical proof that, on our released datasets, training with pseudo-labels yields better performance than directly learning from single-label supervision. In addition, we develop a pseudo-label-based toxicity detection method. Extensive experimental results show that our approach significantly surpasses advanced baselines, including GPT-4o and DeepSeek, thus enabling more accurate and reliable evaluation of multi-label toxicity in LLM-generated content.

Zenghuang Fu, Xiaofeng Han, Mingda Jia et al.

Multi-object tracking (MOT) from unmanned aerial vehicles (UAVs) presents unique challenges due to unpredictable object motion, frequent occlusions, and limited appearance cues inherent to aerial viewpoints. These issues are further exacerbated by abrupt UAV movements, leading to unreliable trajectory estimation and identity switches. Conventional motion models, such as Kalman filters or static sequence encoders, often fall short in capturing both linear and non-linear dynamics under such conditions. To tackle these limitations, we propose DMTrack, a deformable motion tracking framework tailored for UAV-based MOT. Our DMTrack introduces three key components: DeformMamba, a deformable state-space predictor that dynamically aggregates historical motion states for adaptive trajectory modeling; MotionGate, a lightweight gating module that fuses Kalman and Mamba predictions based on motion context and uncertainty; and an uncertainty-aware association strategy that enhances identity preservation by aligning motion trends with prediction confidence. Extensive experiments on the VisDrone-MOT and UAVDT benchmarks demonstrate that our DMTrack achieves state-of-the-art performance in identity consistency and tracking accuracy, particularly under high-speed and non-linear motion. Importantly, our method operates without appearance models and maintains competitive efficiency, highlighting its practicality for robust UAV-based tracking.

Shide Du, Chunming Wu, Zihan Fang et al.

Deep anchor-based multi-view clustering methods enhance the scalability of neural networks by utilizing representative anchors to reduce the computational complexity of large-scale clustering. Despite their scalability advantages, existing approaches often incorporate anchor structures in a heuristic or task-agnostic manner, either through post-hoc graph construction or as auxiliary components for message passing. Such designs overlook the core structural demands of anchor-based clustering, neglecting key optimization principles. To bridge this gap, we revisit the underlying optimization problem of large-scale anchor-based multi-view clustering and unfold its iterative solution into a novel deep network architecture, termed LargeMvC-Net. The proposed model decomposes the anchor-based clustering process into three modules: RepresentModule, NoiseModule, and AnchorModule, corresponding to representation learning, noise suppression, and anchor indicator estimation. Each module is derived by unfolding a step of the original optimization procedure into a dedicated network component, providing structural clarity and optimization traceability. In addition, an unsupervised reconstruction loss aligns each view with the anchor-induced latent space, encouraging consistent clustering structures across views. Extensive experiments on several large-scale multi-view benchmarks show that LargeMvC-Net consistently outperforms state-of-the-art methods in terms of both effectiveness and scalability.

Jiwei Fang, Bin Zhang, Changwei Wang et al.

Verifying the complex and multi-step reasoning of Large Language Models (LLMs) is a critical challenge, as holistic methods often overlook localized flaws. Step-by-step validation is a promising alternative, yet existing methods are often rigid. They struggle to adapt to diverse reasoning structures, from formal proofs to informal natural language narratives. To address this adaptability gap, we propose the Graph of Verification (GoV), a novel framework for adaptable and multi-granular verification. GoV's core innovation is its flexible "node block" architecture. This mechanism allows GoV to adaptively adjust its verification granularity--from atomic steps for formal tasks to entire paragraphs for natural language--to match the native structure of the reasoning process. This flexibility allows GoV to resolve the fundamental trade-off between verification precision and robustness. Experiments on both well-structured and loosely-structured benchmarks demonstrate GoV's versatility. The results show that GoV's adaptive approach significantly outperforms both holistic baselines and other state-of-the-art decomposition-based methods, establishing a new standard for training-free reasoning verification.

Zherui Zhang, Jiaxin Wu, Changwei Wang et al.

Prompt learning as a parameter-efficient method that has been widely adopted to adapt Vision-Language Models (VLMs) to downstream tasks. While hard-prompt design requires domain expertise and iterative optimization, soft-prompt methods rely heavily on task-specific hard labels, limiting their generalization to unseen categories. Recent popular distillation-based prompt learning methods improve generalization by exploiting larger teacher VLMs and unsupervised knowledge transfer, yet their repetitive teacher model online inference sacrifices the inherent training efficiency advantage of prompt learning. In this paper, we propose {\large {\textbf{F}}}aster {\large {\textbf{D}}}istillation-{\large {\textbf{B}}}ased {\large {\textbf{P}}}rompt {\large {\textbf{L}}}earning (\textbf{FDBPL}), which addresses these issues by sharing soft supervision contexts across multiple training stages and implementing accelerated I/O. Furthermore, FDBPL introduces a region-aware prompt learning paradigm with dual positive-negative prompt spaces to fully exploit randomly cropped regions that containing multi-level information. We propose a positive-negative space mutual learning mechanism based on similarity-difference learning, enabling student CLIP models to recognize correct semantics while learning to reject weakly related concepts, thereby improving zero-shot performance. Unlike existing distillation-based prompt learning methods that sacrifice parameter efficiency for generalization, FDBPL maintains dual advantages of parameter efficiency and strong downstream generalization. Comprehensive evaluations across 11 datasets demonstrate superior performance in base-to-new generalization, cross-dataset transfer, and robustness tests, achieving $2.2\times$ faster training speed.

Tian Jin, Enjun Du, Changwei Wang et al.

Parameter-efficient transfer learning (PETL) aims to reduce the scales of pretrained models for multiple downstream tasks. However, as the models keep scaling up, the memory footprint of existing PETL methods is not significantly reduced compared to the reduction of learnable parameters. This limitation hinders the practical deployment of PETL methods on memory-constrained devices. To this end, we proposed a new PETL framework, called Structure to Activation (S2A), to reduce the memory footprint of activation during fine-tuning. Specifically, our framework consists of: 1) Activation modules design(i.e., bias, prompt and side modules) in the parametric model structure, which results in a significant reduction of adjustable parameters and activation memory; 2) 4-bit quantization of activations based on their derivatives for non-parametric structures (e.g., nonlinear functions), which maintains accuracy while significantly reducing memory usage. Our S2A method consequently offers a lightweight solution in terms of both parameters and memory footprint. We evaluated S2A with different backbones and performed extensive experiments on various datasets to evaluate the effectiveness. The results show that our methods not only outperform existing PETL techniques, achieving a fourfold reduction in GPU memory footprint on average, but also shows competitive performance in accuracy with fewer tunable parameters. These demonstrate that our method is highly suitable for practical transfer learning on hardware-constrained devices.

Zhengpeng Xie, Jiahang Cao, Changwei Wang et al.

In this paper, we argue that mutual distillation between reinforcement learning policies serves as an implicit regularization, preventing them from overfitting to irrelevant features. We highlight two separate contributions: (i) Theoretically, for the first time, we prove that enhancing the policy robustness to irrelevant features leads to improved generalization performance. (ii) Empirically, we demonstrate that mutual distillation between policies contributes to such robustness, enabling the spontaneous emergence of invariant representations over pixel inputs. Ultimately, we do not claim to achieve state-of-the-art performance but rather focus on uncovering the underlying principles of generalization and deepening our understanding of its mechanisms.

Yu Zhang, Qi Zhang, Zixuan Gong et al.

Contrastive Language-Image Pretraining (CLIP) has achieved remarkable success, leading to rapid advancements in multimodal studies. However, CLIP faces a notable challenge in terms of inefficient data utilization. It relies on a single contrastive supervision for each image-text pair during representation learning, disregarding a substantial amount of valuable information that could offer richer supervision. Additionally, the retention of non-informative tokens leads to increased computational demands and time costs, particularly in CLIP's ViT image encoder. To address these issues, we propose Multi-Perspective Language-Image Pretraining (MLIP). In MLIP, we leverage the frequency transform's sensitivity to both high and low-frequency variations, which complements the spatial domain's sensitivity limited to low-frequency variations only. By incorporating frequency transforms and token-level alignment, we expand CILP's single supervision into multi-domain and multi-level supervision, enabling a more thorough exploration of informative image features. Additionally, we introduce a token merging method guided by comprehensive semantics from the frequency and spatial domains. This allows us to merge tokens to multi-granularity tokens with a controllable compression rate to accelerate CLIP. Extensive experiments validate the effectiveness of our design.

Changwei Wang, Rongtao Xu, Shibiao Xu et al.

Accurate lung lesion segmentation from Computed Tomography (CT) images is crucial to the analysis and diagnosis of lung diseases such as COVID-19 and lung cancer. However, the smallness and variety of lung nodules and the lack of high-quality labeling make the accurate lung nodule segmentation difficult. To address these issues, we first introduce a novel segmentation mask named Soft Mask which has richer and more accurate edge details description and better visualization and develop a universal automatic Soft Mask annotation pipeline to deal with different datasets correspondingly. Then, a novel Network with detailed representation transfer and Soft Mask supervision (DSNet) is proposed to process the input low-resolution images of lung nodules into high-quality segmentation results. Our DSNet contains a special Detail Representation Transfer Module (DRTM) for reconstructing the detailed representation to alleviate the small size of lung nodules images, and an adversarial training framework with Soft Mask for further improving the accuracy of segmentation. Extensive experiments validate that our DSNet outperforms other state-of-the-art methods for accurate lung nodule segmentation and has strong generalization ability in other accurate medical segmentation tasks with competitive results. Besides, we provide a new challenging lung nodules segmentation dataset for further studies.