Saved in:
Bibliographic Details
Main Authors: Chen, Boyu, Yue, Zhengrong, Chen, Siran, Wang, Zikang, Liu, Yang, Li, Peng, Wang, Yali
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2503.10200
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911328595083264
author Chen, Boyu
Yue, Zhengrong
Chen, Siran
Wang, Zikang
Liu, Yang
Li, Peng
Wang, Yali
author_facet Chen, Boyu
Yue, Zhengrong
Chen, Siran
Wang, Zikang
Liu, Yang
Li, Peng
Wang, Yali
contents Existing MLLMs encounter significant challenges in modeling the temporal context within long videos. Currently, mainstream Agent-based methods use external tools to assist a single MLLM in answering long video questions. Despite such tool-based support, a solitary MLLM still offers only a partial understanding of long videos, resulting in limited performance. In order to better address long video tasks, we introduce LVAgent, the first framework enabling multi-round dynamic collaboration of MLLM agents in long video understanding. Our method consists of four key steps: 1) Selection: We pre-select appropriate agents from the model library to form optimal agent teams based on different tasks. 2) Perception: We design an effective retrieval scheme for long videos to improve the coverage of critical temporal segments while maintaining computational efficiency. 3) Action: Agents answer long video questions and exchange reasons. 4) Reflection: We evaluate each agent's performance in each round of discussion and optimize the agent team for dynamic collaboration. The agents iteratively refine their answers by multi-round dynamical collaboration of MLLM agents. LVAgent is the first agent system method that outperforms all closed-source models (like GPT-4o) and open-source models (like InternVL-2.5 and Qwen2-VL) in the long video understanding tasks. Our LVAgent achieves an accuracy of 80\% on four mainstream long video understanding tasks. Notably, LVAgent improves accuracy by 13.3\% on LongVideoBench. Code is available at https://github.com/64327069/LVAgent.
format Preprint
id arxiv_https___arxiv_org_abs_2503_10200
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle LVAgent: Long Video Understanding by Multi-Round Dynamical Collaboration of MLLM Agents
Chen, Boyu
Yue, Zhengrong
Chen, Siran
Wang, Zikang
Liu, Yang
Li, Peng
Wang, Yali
Computer Vision and Pattern Recognition
Existing MLLMs encounter significant challenges in modeling the temporal context within long videos. Currently, mainstream Agent-based methods use external tools to assist a single MLLM in answering long video questions. Despite such tool-based support, a solitary MLLM still offers only a partial understanding of long videos, resulting in limited performance. In order to better address long video tasks, we introduce LVAgent, the first framework enabling multi-round dynamic collaboration of MLLM agents in long video understanding. Our method consists of four key steps: 1) Selection: We pre-select appropriate agents from the model library to form optimal agent teams based on different tasks. 2) Perception: We design an effective retrieval scheme for long videos to improve the coverage of critical temporal segments while maintaining computational efficiency. 3) Action: Agents answer long video questions and exchange reasons. 4) Reflection: We evaluate each agent's performance in each round of discussion and optimize the agent team for dynamic collaboration. The agents iteratively refine their answers by multi-round dynamical collaboration of MLLM agents. LVAgent is the first agent system method that outperforms all closed-source models (like GPT-4o) and open-source models (like InternVL-2.5 and Qwen2-VL) in the long video understanding tasks. Our LVAgent achieves an accuracy of 80\% on four mainstream long video understanding tasks. Notably, LVAgent improves accuracy by 13.3\% on LongVideoBench. Code is available at https://github.com/64327069/LVAgent.
title LVAgent: Long Video Understanding by Multi-Round Dynamical Collaboration of MLLM Agents
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2503.10200