Saved in:
Bibliographic Details
Main Authors: Yang, Zhongyu, Yang, Zuhao, Zhan, Shuo, Yue, Tan, Pang, Wei, Yuan, Yingfang
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2604.05079
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913009309319168
author Yang, Zhongyu
Yang, Zuhao
Zhan, Shuo
Yue, Tan
Pang, Wei
Yuan, Yingfang
author_facet Yang, Zhongyu
Yang, Zuhao
Zhan, Shuo
Yue, Tan
Pang, Wei
Yuan, Yingfang
contents Video question answering (VideoQA) is a challenging task that requires integrating spatial, temporal, and semantic information to capture the complex dynamics of video sequences. Although recent advances have introduced various approaches for video understanding, most existing methods still rely on locating relevant frames to answer questions rather than reasoning through the evolving storyline as humans do. Humans naturally interpret videos through coherent storylines, an ability that is crucial for making robust and contextually grounded predictions. To address this gap, we propose SVAgent, a storyline-guided cross-modal multi-agent framework for VideoQA. The storyline agent progressively constructs a narrative representation based on frames suggested by a refinement suggestion agent that analyzes historical failures. In addition, cross-modal decision agents independently predict answers from visual and textual modalities under the guidance of the evolving storyline. Their outputs are then evaluated by a meta-agent to align cross-modal predictions and enhance reasoning robustness and answer consistency. Experimental results demonstrate that SVAgent achieves superior performance and interpretability by emulating human-like storyline reasoning in video understanding.
format Preprint
id arxiv_https___arxiv_org_abs_2604_05079
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle SVAgent: Storyline-Guided Long Video Understanding via Cross-Modal Multi-Agent Collaboration
Yang, Zhongyu
Yang, Zuhao
Zhan, Shuo
Yue, Tan
Pang, Wei
Yuan, Yingfang
Computer Vision and Pattern Recognition
Video question answering (VideoQA) is a challenging task that requires integrating spatial, temporal, and semantic information to capture the complex dynamics of video sequences. Although recent advances have introduced various approaches for video understanding, most existing methods still rely on locating relevant frames to answer questions rather than reasoning through the evolving storyline as humans do. Humans naturally interpret videos through coherent storylines, an ability that is crucial for making robust and contextually grounded predictions. To address this gap, we propose SVAgent, a storyline-guided cross-modal multi-agent framework for VideoQA. The storyline agent progressively constructs a narrative representation based on frames suggested by a refinement suggestion agent that analyzes historical failures. In addition, cross-modal decision agents independently predict answers from visual and textual modalities under the guidance of the evolving storyline. Their outputs are then evaluated by a meta-agent to align cross-modal predictions and enhance reasoning robustness and answer consistency. Experimental results demonstrate that SVAgent achieves superior performance and interpretability by emulating human-like storyline reasoning in video understanding.
title SVAgent: Storyline-Guided Long Video Understanding via Cross-Modal Multi-Agent Collaboration
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2604.05079