ConceptAttention: Diffusion Transformers Learn Highly Interpretable Features
机构 * ibm（IBM研究院）
Comments Oral Presentation at ICML 2025, Best Paper Award at CVPR Workshop on Visual Concepts
期刊&会议
CVPR
Conference on Computer Vision and Pattern Recognition · 会议 · Computer Vision
NN-Former: Rethinking Graph Structure in Neural Architecture Representation
机构 * State Key Laboratory of Multimedia Information Processing, School of Computer Science, Peking University（多媒体信息处理国家重点实验室，计算机学院，北京大学） ； Northwestern Polytechnical University（西北工业大学） ； Harbin Institute of Technology (Shenzhen)（哈尔滨工业大学（深圳））
Comments Accepted to CVPR 2025. Code is avaiable at https://github.com/XuRuihan/NNFormer
Instant Particle Size Distribution Measurement Using CNNs Trained on Synthetic Data
机构 * College of Computing, University Mohammed VI Polytechnic（穆拉比特大学第六理工学院）
Comments Accepted at the Synthetic Data for Computer Vision Workshop @ CVPR 2025. 10 pages, 5 figures. Code available at https://github.com/YasserElj/Synthetic-Granular-Gen
Moment Sampling in Video LLMs for Long-Form Video QA
机构 * University of Massachusetts Amherst（马萨诸塞大学阿默斯特分校） ； Dolby Laboratories（杜比实验室）
Comments Workshop on Video Large Language Models (VidLLMs) at CVPR 2025
Instruct-4DGS: Efficient Dynamic Scene Editing via 4D Gaussian-based Static-Dynamic Separation
机构 * DGIST ； KAIST（韩国国立科学技术院）
Comments Accepted to CVPR 2025. The first two authors contributed equally
Revisiting Audio-Visual Segmentation with Vision-Centric Transformer
机构 * Hefei University of Technology（合肥工业大学） ； Chinese Academy of Sciences（中国科学院） ； Beihang University（北航） ； Sangfor Technologies（深信服技术）
Comments Accepted by CVPR 2025; Code: https://github.com/spyflying/VCT_AVS; Models: https://huggingface.co/nowherespyfly/VCT_AVS
MTADiffusion: Mask Text Alignment Diffusion Model for Object Inpainting
机构 * MT Lab, Meitu Inc（美图实验室，美图公司） ； National University of Singapore（新加坡国立大学） ； Department of Computer Science and Technology, BNRist, IDG/McGovern Institute for Brain Research, Tsinghua University（计算机科学与技术系，BNRist，IDG/McGovern脑科学研究院，清华大学）
Comments CVPR 2025
CountLLM: Towards Generalizable Repetitive Action Counting via Large Language Model
机构 * Peking University（北京大学） ； University of Washington（华盛顿大学） ； University of Copenhagen（哥本哈根大学）
Comments Accepted by CVPR 2025
Analyzing and Fine-Tuning Whisper Models for Multilingual Pilot Speech Transcription in the Cockpit
机构 * Institute of Data Science（数据科学研究所） ； German Aerospace Center（德国航空航天中心）
Comments Computer Vision and Pattern Recognition (CVPR) 2025 Workshops
SceneDiffuser++: City-Scale Traffic Simulation via a Generative World Model
机构 * Waymo LLC（Waymo公司） ； UT Austin（德克萨斯大学奥斯汀分校）
Comments Accepted to CVPR 2025
Mr. DETR++: Instructive Multi-Route Training for Detection Transformers with Mixture-of-Experts
机构 * The University of Hong Kong（香港大学） ； Meituan Inc.（美团公司）
Comments Under review. Extended version of our CVPR 2025 paper, see arXiv:2412.10028v3
Distilling Normalizing Flows
机构 * University of Oregon（俄勒冈大学） ； HSE University（俄罗斯高等经济大学） ； Picsart AI Research（Picsart人工智能研究所）
Comments Published in eLVM @ CVPR (https://openaccess.thecvf.com/content/CVPR2025W/eLVM/html/Walton_Distilling_Normalizing_Flows_CVPRW_2025_paper)
High Temporal Consistency through Semantic Similarity Propagation in Semi-Supervised Video Semantic Segmentation for Autonomous Flight
机构 * Télécom Paris, Institut Polytechnique de Paris（巴黎电信学院，巴黎理工大学） ； Fraunhofer IVI（弗劳恩霍夫IVI研究所）
Comments Accepted by CVPR2025
Journal ref Proceedings of the Computer Vision and Pattern Recognition Conference (CVPR), 2025, pp. 1461-1471
Efficient Image Generation with Variadic Attention Heads
机构 * University of Oregon（俄勒冈大学） ； University of Texas at Austin（德克萨斯大学奥斯汀分校）
Comments Published in eLVM @ CVPR (https://openaccess.thecvf.com/content/CVPR2025W/eLVM/html/Walton_Efficient_Image_Generation_with_Variadic_Attention_Heads_CVPRW_2025_paper) | Formerly named StyleNAT: Giving Each Head a New Perspective |
Joint attitude estimation and 3D neural reconstruction of non-cooperative space objects
机构 * CS Group（计算机科学组） ； CNES（国家空间科学中心）
Comments accepted for CVPR 2025 NFBCC workshop
KD-DETR: Knowledge Distillation for Detection Transformer with Consistent Distillation Points Sampling
机构 * Baidu VIS（百度视觉）
Comments Accepted to CVPR 2024
SceneCrafter: Controllable Multi-View Driving Scene Editing
机构 * Waymo ； University of Texas at Austin（德克萨斯大学奥斯汀分校） ； Johns Hopkins University（约翰霍普金斯大学） ； Google DeepMind（谷歌DeepMind）
Comments CVPR 2025
Emergence of Text Readability in Vision Language Models
机构 * Naver AI Lab ； Seoul National University（首尔国立大学） ； Computer Vision Laboratory, ECE（计算机视觉实验室，电子与计算机工程系） ； IPAI
Comments EVAL-FoMo Workshop @ CVPR 2025
MedErr-CT: A Visual Question Answering Benchmark for Identifying and Correcting Errors in CT Reports
机构 * Department of Biomedical Engineering, University of Ulsan College of Medicine（生物医学工程系，釜山大学医学院）
Comments 14 pages, 5 figures, submitted to CVPR 2025
RareSpot: Spotting Small and Rare Wildlife in Aerial Imagery with Multi-Scale Consistency and Context-Aware Augmentation
机构 * University of California, Santa Barbara（加州大学圣芭芭拉分校） ； Smithsonian National Zoo and Conservation Biology Institute（史密森尼国家动物园与保护生物学研究所） ； Stanford University（斯坦福大学）
Comments Accepted to the CVPR 2025 Workshop on Computer Vision for Animal Behavior Tracking and Modeling (CV4Animals)
Object-aware Sound Source Localization via Audio-Visual Scene Understanding
机构 * Kyung Hee University（庆熙大学） ； KAIST AI（韩国科学技术院人工智能研究所） ； Sungkyunkwan University（成均馆大学）
Comments Accepted at CVPR 2025
Journal ref Proceedings of the Computer Vision and Pattern Recognition Conference (CVPR), 2025, pp. 8342-8351
Rethinking Decoder Design: Improving Biomarker Segmentation Using Depth-to-Space Restoration and Residual Linear Attention
机构 * School of Computing, KAIST, Republic of Korea（韩国科学技术院计算机学院）
Comments Proceedings of the Computer Vision and Pattern Recognition Conference (CVPR), 2025, pp. 30861-30871
GCE-Pose: Global Context Enhancement for Category-level Object Pose Estimation
机构 * Technical University of Munich（慕尼黑技术大学） ； Munich Center for Machine Learning（慕尼黑机器学习中心） ； XYZ Robotics（XYZ机器人）
Comments CVPR 2025 accepted
Low-power, Continuous Remote Behavioral Localization with Event Cameras
Comments 13 pages, 8 figures, 12 tables, Project page: https://tub-rip.github.io/eventpenguins/
Journal ref IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, 2024
CDI: Copyrighted Data Identification in Diffusion Models
机构 * Warsaw University of Technology, IDEAS NCBR（华沙理工大学，IDEAS NCBR） ； CISPA Helmholtz Center for Information Security（CISPA 河岸信息安全中心）
Comments Accepted at CVPR2025 (Conference on Computer Vision and Pattern Recognition) Code available at https://github.com/sprintml/copyrighted_data_identification
Relation3D: Enhancing Relation Modeling for Point Cloud Instance Segmentation
机构 * University of Science and Technology of China（中国科学技术大学）
Comments Accepted by CVPR 2025. Code: https://github.com/Howard-coder191/Relation3D
HIRE: Lightweight High-Resolution Image Feature Enrichment for Multimodal LLMs
机构 * Media and Data Science Research Lab, Adobe（Adobe媒体与数据科学研究实验室）
Comments Accepted in CVPR 2025 Workshop on What's Next in Multimodal Foundational Models
ILIAS: Instance-Level Image retrieval At Scale
Comments CVPR 2025
XYScanNet: A State Space Model for Single Image Deblurring
机构 * Texas A&M University（德克萨斯A&M大学） ； Johns Hopkins University（约翰霍普金斯大学）
Journal ref Proceedings of the Computer Vision and Pattern Recognition Conference (CVPR) Workshops, 2025
G3Flow: Generative 3D Semantic Flow for Pose-aware and Generalizable Object Manipulation
机构 * The University of Hong Kong（香港大学） ； Institute of Artificial Intelligence (TeleAI), China Telecom（人工智能研究所（TeleAI），中国电信） ； Shenzhen University（深圳大学） ； AgileX Robotics（AgileX机器人） ； GDIIST ； HKU Shanghai Intelligent Computing Research Center（香港大学上海智能计算研究中心）
Comments Webpage: https://tianxingchen.github.io/G3Flow/, accepted to CVPR 2025
