CoMPaSS: Enhancing Spatial Understanding in Text-to-Image Diffusion Models
机构 * Zhejiang University（浙江大学） ； vivo ； Ant Group（蚂蚁集团）
Comments 21 pages, 12 figures. Accepted to ICCV 2025
期刊&会议
ICCV
International Conference on Computer Vision · 会议 · Computer Vision
Diffusion-Based Data Augmentation for Medical Image Segmentation
机构 * School of Computer Science and Engineering, Beihang University（北京航空航天大学计算机科学与工程学院） ； Dept. of Engineering for Innovation Medicine, University of Verona（威尼斯大学创新医学工程系）
Comments Accepted to CVAMD Workshop at ICCV 2025
A Contrastive Learning-Guided Confident Meta-learning for Zero Shot Anomaly Detection
机构 * Dept. of Engineering for Innovation Medicine, University of Verona（创新医学工程系，威尼斯大学） ； Faculty of Computer and Information Science, University of Ljubljana（计算机与信息科学系，卢布尔雅那大学） ； Qualyco S.r.l.（Qualyco公司）
Comments Accepted to VISION Workshop at ICCV 2025
TemCoCo: Temporally Consistent Multi-modal Video Fusion with Visual-Semantic Collaboration
机构 * Electronic Information School, Wuhan University, Wuhan 430072, China（武汉大学电子信息学院）
Comments Accepted by ICCV 2025
Robust Anomaly Detection in Industrial Environments via Meta-Learning
机构 * Dept. of Engineering for Innovation Medicine, University of Verona（创新医学工程系，威尼斯大学） ； Qualyco S.r.l.（Qualyco公司）
Comments Accepted to VISION Workshop at ICCV 2025
Few-Shot Pattern Detection via Template Matching and Regression
机构 * Pohang University of Science and Technology (POSTECH)（浦项科学技术大学）
Comments Accepted to ICCV 2025 (highlight)
TinyGiantVLM: A Lightweight Vision-Language Architecture for Spatial Reasoning under Resource Constraints
机构 * University of Science, VNU-HCM（越南胡志明市国家大学） ； Vietnam National University（越南国家大学）
Comments Accepted for presentation at the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, 2025
Journal ref IEEE/CVF International Conference on Computer Vision (ICCV) Workshops, Hawaii, 2025
DanceEditor: Towards Iterative Editable Music-driven Dance Generation with Open-Vocabulary Descriptions
机构 * Peking University（北京大学） ； The Hong Kong University of Science and Technology（香港科技大学） ； Beijing University of Posts and Telecommunications（北京邮电大学）
Journal ref ICCV 2025
OmniCache: A Trajectory-Oriented Global Perspective on Training-Free Cache Reuse for Diffusion Transformer Models
机构 * Zhipu AI（智谱AI）
Comments Accepted by ICCV 2025
HyTIP: Hybrid Temporal Information Propagation for Masked Conditional Residual Video Coding
Comments Accepted to ICCV 2025
MemoryTalker: Personalized Speech-Driven 3D Facial Animation via Audio-Guided Stylization
机构 * Chung-Ang University（Chung-Ang 大学） ； Korea University（韩国大学）
Comments Accepted in ICCV 2025; Project Page: https://cau-irislab.github.io/ICCV25-MemoryTalker/
Taming the Untamed: Graph-Based Knowledge Retrieval and Reasoning for MLLMs to Conquer the Unknown
机构 * Osaka University（大阪大学） ； Mitsubishi Electric Corp.（三菱电机公司）
Comments Aligned with ICCV 2025 camera-ready version
Beyond Blur: A Fluid Perspective on Generative Diffusion Models
Comments ICCV 2025 main conference, 8 pages paper, 20 pages appendix, 24 figures, supplementary pseudocode in appendix, https://iccv.thecvf.com/virtual/2025/poster/1176
Seeing the Trees for the Forest: Rethinking Weakly-Supervised Medical Visual Grounding
机构 * Australian Institute for Machine Learning, University of Adelaide（澳大利亚机器学习研究所，阿德莱德大学） ； Mohamed bin Zayed University of Artificial Intelligence（穆罕默德·本·扎耶德人工智能大学） ； Macquarie University（麦考瑞大学） ； Hanoi University of Science and Technology（河内科学技术大学） ； University of Wollongong（沃林根大学） ； Flinders University（弗林德斯大学）
Comments Accepted at ICCV 2025 (Highlight)
GigaTok: Scaling Visual Tokenizers to 3 Billion Parameters for Autoregressive Image Generation
机构 * The University of Hong Kong（香港大学） ； ByteDance Seed Project（字节跳动种子项目）
Comments ICCV 2025. Project page: https://silentview.github.io/GigaTok
GaussianFlowOcc: Sparse and Weakly Supervised Occupancy Estimation using Gaussian Splatting and Temporal Flow
机构 * Robert Bosch GmbH（罗伯特·博世有限公司） ； University of Münster（穆斯尔大学）
Comments Accepted to ICCV 2025
Orchid: Image Latent Diffusion for Joint Appearance and Geometry Generation
机构 * Google DeepMind（谷歌DeepMind） ； Georgia Institute of Technology（佐治亚理工学院） ； Waymo
Comments Accepted to ICCV 2025. Project webpage: https://orchid3d.github.io
Addressing Text Embedding Leakage in Diffusion-based Image Editing
机构 * Graduate School of AI, POSTECH（POSTECH人工智能研究生院） ； Dept. of CSE, POSTECH（POSTECH计算机科学与工程系）
Comments Accepted to ICCV 2025
Leveraging the Power of MLLMs for Gloss-Free Sign Language Translation
机构 * Department of Artificial Intelligence, Yonsei University（人工智能系，延世大学） ； Graduate School of Information, Yonsei University（信息研究生院，延世大学）
Comments Accepted by ICCV 2025
TokenUnify: Scaling Up Autoregressive Pretraining for Neuron Segmentation
机构 * MoE Key Laboratory of Brain-inspired Intelligent Perception and Cognition, University of Science and Technology of China（脑启发智能感知与认知大学科学与技术大学实验室） ； Institute of Artificial Intelligence, Hefei Comprehensive National Science Center（人工智能研究所） ； Institute for Brain and Intelligence, Fudan University（脑与智能研究所）
Comments Accepted by ICCV 2025
MDD: A Dataset for Text-and-Music Conditioned Duet Dance Generation
机构 * Purdue University（普渡大学）
Comments Accepted at ICCV 2025. Project page: https://gprerit96.github.io/mdd-page
AIM 2025 Low-light RAW Video Denoising Challenge: Dataset, Methods and Results
机构 * AIM 2025 Low-light RAW Video Denoising Challenge（AIM 2025 低光照RAW视频去噪挑战）
Comments Challenge report from Advances in Image Manipulation workshop held at ICCV 2025
Toward Socially Aware Vision-Language Models: Evaluating Cultural Competence Through Multimodal Story Generation
机构 * KIIT Deemed University（KIIT大学） ； Indian Institute of Technology (IIT) Bhubaneswar（印度理工学院（Bhubaneswar分校））
Comments Accepted at ASI @ ICCV 2025
