Event-Anchored Frame Selection for Effective Long-Video Understanding
Key Laboratory of Multimedia Trusted Perception and Efficient Computing, Ministry of Education of China, Xiamen University(教育部多媒体可信感知与高效计算重点实验室,厦门大学);College of Computer and Data Science, Fuzhou University(福州大学计算机与数据科学学院)
Wang Chen, Yongdong Luo, Yuhui Zeng, Luojun Lin, Tianyu Xie, Fei Chao, Rongrong Ji, Xiawu Zheng
Massive frame redundancy and limited context window make efficient frame selection crucial for long-video understanding with large vision-language models (LVLMs). Prevailing approaches, however, adopt a flat sampling paradigm which treats the video as an unstructured collection of frames. In this paper, we introduce Event-Anchored Frame Selection (EFS), a hierarchical, event-aware pipeline. Leveraging self-supervised DINO embeddings, EFS first partitions the video stream into visually homogeneous temporal segments, which serve as proxies for semantic events. Within each event, it then selects the most query-relevant frame as an anchor. These anchors act as structural priors that guide a global refinement stage using an adaptive Maximal Marginal Relevance (MMR) scheme. This pipeline ensures the final keyframe set jointly optimizes for event coverage, query relevance, and visual diversity. As a training-free, plug-and-play module, EFS can be seamlessly integrated into off-the-shelf LVLMs, yielding substantial gains on challenging video understanding benchmarks. Specifically, when applied to LLaVA-Video-7B, EFS improves accuracy by 4.7%, 4.9%, and 8.8% on VideoMME, LongVideoBench, and MLVU, respectively.
AV-GRPO: Modality-Anchored Decoupling Diffusion Reinforcement Learning for Joint Audio-Video Generation
Shanghai AI Laboratory(上海人工智能实验室);The Hong Kong Polytechnic University(香港理工大学);National University of Singapore(新加坡国立大学);Nanyang Technological University(南洋理工大学);Zhejiang University(浙江大学)
Sparrow: Text-Anchored Window Attention with Visual-Semantic Glimpsing for Speculative Decoding in Video LLMs
National Key Laboratory of Parallel and Distributed Computing(平行与分布式计算国家重点实验室);College of Computer Science and Technology(计算机科学与技术学院);National University of Defense Technology(国防科技大学)
EgoGenesis: Egocentric World-Action Modeling with Online Anchored Projective Memory and Action-3D RoPE
Shanghai Jiao Tong University(上海交通大学);Alibaba Group(阿里巴巴集团);Tianji KernalMind Co., Ltd.(天机芯智有限公司);The Hong Kong University of Science and Technology(香港科技大学);Southeast University(东南大学);Renmin University of China(中国人民大学);The University of Tokyo(东京大学)
RiskProp: Collision-Anchored Self-Supervised Risk Propagation for Early Accident Anticipation
School of Computer Science, Wuhan University(武汉大学计算机学院);Zhongguancun Academy(中关村学院);Didi Chuxing(滴滴出行);University of Science and Technology of China(中国科学技术大学)
Prior-Anchored Debiasing for Long-Tailed Multi-Organ Pathology Report Generation
City University of Hong Kong, Hong Kong SAR(香港城市大学);University of Nottingham, United Kingdom(诺丁汉大学);University of Massachusetts Boston, United States(马萨诸塞大学波士顿分校)
GeoFF3D: Coordinate-Anchored Feed-Forward Reconstruction for Large-Scale UAV Mapping
Central South University(中南大学);Hunan Engineering Research Center of 3D Real Scene Construction and Application Technology(湖南省3D真实场景构建与应用技术工程研究中心);National University of Defense Technology(国防科技大学)