跳到正文
arXiv · AI· Dian Jiao, Yuqian Yuan, Jiaqi Zhu, Mingjian Gao, Siwen Tan, Yunqi Cao, Jinhui Li, Sijing Li, Zhenxuan Fan, Haoyang Wang, Wenqiao Zhang, Siliang Tang·· 11 天前AI 评分20

V-STORM:面向忠实细粒度时空视频推理的结构感知 Grounded Chain of Thought 数据集

V-STORM: A Structure-Aware Grounded Chain of Thought Dataset for Faithful and Fine-Grained Spatio-Temporal Video Reasoning

AI 导读

V-STORM 将视频转化为 Spatio-Temporal Scene Graphs 并推导 Grounded Chain of Thought,构建了 260K 训练集和 3K 人工验证 benchmark。

来源:arXiv · AI · arxiv.org