Lune

CVPR2025Top-tier venue

DynFocus: Dynamic Cooperative Network Empowers LLMs with Video Understanding

Yudong Han, Qingpei Guo, Liyuan Pan, Liu Liu, Yu Guan, Ming Yang

Abstract

The challenge in LLM-based video understanding lies in preserving visual and semantic information in long videos while maintaining a memory-affordable token count. However, redundancy and correspondence in videos have hindered the performance potential of existing methods. Through statistical learning on current datasets, we observe that redundancy occurs in both repeated and answerirrelevant frames, and the corresponding frames vary with different questions. This suggests the possibility of adopting dynamic encoding to balance detailed video information preservation with token budget reduction. To this end, we propose a dynamic cooperative network, DynFocus, for memory-efficient video encoding in this paper. Specifically, i) a Dynamic Event Prototype Estimation (DPE) module to dynamically select meaningful frames for question answering; (ii) a Compact Cooperative Encoding (CCE) module that encodes meaningful frames with detailed visual appearance and the remaining frames with sketchy perception separately. We evaluate our method on five publicly available benchmarks, and experimental results consistently demonstrate that our method achieves competitive performance. Code is available at https://github.com/Simon98-AI/ DynFocus

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

5 top-tier papers have cited it since. Know what they found before your reviewers do.

Questions to start from

Lune can draw on

the full paper · the top-tier papers that cite it · related work

Get the answer

Free to start. No credit card required.

Lineage

  • InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning cites Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answering
  • EmoThinker: Advancing Visual-Acoustic Emotion Analysis via Structural Token Selection and Chain-of-Thought Reasoning cites Storyboard-guided Alignment for Fine-grained Video Action Recognition
  • Storyboard-guided Alignment for Fine-grained Video Action Recognition cites Learning Transferable Visual Models From Natural Language Supervision
  • A Multi-Agent Perception-Action Alliance for Efficient Long Video Reasoning cites Learning Transferable Visual Models From Natural Language Supervision
  • A Multi-Agent Perception-Action Alliance for Efficient Long Video Reasoning cites Deep Video Discovery: Agentic Search with Tool Use for Long-form Video Understanding
  • Frozen in Time: A Joint Video and Image Encoder for End-to-End Retrieval cites Learning Transferable Visual Models From Natural Language Supervision
  • Question-guided Visual Compression with Memory Feedback for Long-Term Video Understanding cites InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning
Builds onCited byLeft to right by year; the more cited, the higher.

Cited by top-tier papers5

  1. Deep Video Discovery: Agentic Search with Tool Use for Long-form Video UnderstandingXiaoyi Zhang, Zhaoyang Jia, Zongyu Guo, Jiahao Li et al.NeurIPS 2025 · 95 citations
  2. Storyboard-guided Alignment for Fine-grained Video Action RecognitionEnqi Liu, Liyuan Pan, Yan Yang, Yiran Zhong et al.NeurIPS 2025 · 3 citations
  3. Question-guided Visual Compression with Memory Feedback for Long-Term Video UnderstandingSosuke Yamao, Natsuki Miyahara, Yuankai Qi, Shun TakeuchiCVPR 2026 · 3 citations
  4. A Multi-Agent Perception-Action Alliance for Efficient Long Video ReasoningYichang Xu, Gaowen Liu, Ramana Rao Kompella, Tiansheng Huang et al.CVPR 2026 · 2 citations
  5. EmoThinker: Advancing Visual-Acoustic Emotion Analysis via Structural Token Selection and Chain-of-Thought ReasoningQinfu Xu, Liyuan Pan, Yiwei Wei, Shaozu Yuan et al.CVPR 2026

Builds on16

  1. Learning Transferable Visual Models From Natural Language SupervisionAlec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh et al.ICML 2021 · 47,906 citations
  2. InstructBLIP: Towards General-purpose Vision-Language Models with Instruction TuningWenliang Dai, Junnan Li, Dongxu Li, Anthony Meng Huat Tiong et al.NeurIPS 2023 · 4,013 citations
  3. Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question AnsweringPan Lu, Swaroop Mishra, Tanglin Xia, Liang Qiu et al.NeurIPS 2022 · 2,727 citations
  4. Frozen in Time: A Joint Video and Image Encoder for End-to-End RetrievalMax Bain, Arsha Nagrani, Gül Varol, Andrew ZissermanICCV 2021 · 1,550 citations
  5. CLEVRER: Collision Events for Video Representation and ReasoningKexin Yi, Chuang Gan, Yunzhu Li, Pushmeet Kohli et al.ICLR 2020 · 584 citations

Related papers

  1. LongVU: Spatiotemporal Adaptive Compression for Long Video-Language UnderstandingXiaoqian Shen, Yunyang Xiong, Changsheng Zhao, Lemeng Wu et al.ICML 2025
  2. FastVID: Dynamic Density Pruning for Fast Video Large Language ModelsLeqi Shen, Guoqiang Gong, Tao He, Yifeng Zhang et al.NeurIPS 2025 · 56 citations
  3. Streaming Long Video Understanding with Large Language ModelsRui Qian, Xiaoyi Dong, Pan Zhang, Yuhang Zang et al.NeurIPS 2024 · 216 citations
  4. VQToken: Neural Discrete Token Representation Learning for Extreme Token Reduction in Video Large Language ModelsHaichao Zhang, Yun FuNeurIPS 2025 · 15 citations
  5. FlexiVideo: Variation-Aware Temporal Dynamics Modeling for Efficient Video UnderstandingDa Peng, Xuesong Yang, Zonghao Guo, Yichen Zhang et al.CVPR 2026
Dusk over the sea between two cliffs drawn in fine vertical lines