精确定位接地视频问答的触发时刻：增强多模态大语言模型中的时空接地
Pinpointing Trigger Moment for Grounded Video QA: Enhancing Spatio-temporal Grounding in Multimodal Large Language Models
AI总结：
针对ICCV 2025 GVQA任务，提出三阶段流水线框架，核心引入基于CORTEX提示的触发时刻以定位目标最清晰帧作为锚点，实现0.4968的HOTA分数，显著超越去年成绩。
摘要：
在本技术报告中，我们介绍了一个用于解决ICCV 2025感知测试挑战赛接地视频问答（GVQA）任务的框架。GVQA任务要求具备鲁棒性的多模态模型，能够对视频内容进行复杂推理，将所得答案在视觉上接地，并在时间上追踪被引用的对象。为实现这一能力，我们提出的方法将GVQA任务分解为三阶段流水线：(1) 视频推理与问答，(2) 时空接地，以及 (3) 追踪。我们的核心贡献是引入了触发时刻，该时刻源自我们提出的CORTEX提示，能够精确定位目标对象最清晰可见的单帧，作为接地和追踪的鲁棒锚点。为此，我们取得了0.4968的HOTA分数，标志着相较于上一年度GVQA任务0.2704的获胜分数有了显著提升。
英文摘要：
In this technical report, we introduce a framework to address Grounded Video Question Answering (GVQA) task for the ICCV 2025 Perception Test Challenge. The GVQA task demands robust multimodal models capable of complex reasoning over video content, grounding the resulting answers visually, and tracking the referenced objects temporally. To achieve this capability, our proposed approach decomposes the GVQA task into a three-stage pipeline: (1) Video Reasoning \& QA, (2) Spatio-temporal Grounding and (3) Tracking. Our key contribution is the introduction of a trigger moment, derived from our proposed CORTEX prompt, which pinpoints the single most visible frame of a target object to serve as a robust anchor for grounding and tracking. To this end, we achieve the HOTA score of 0.4968, which marks a significant improvement over the previous year's winning score of 0.2704 on GVQA task.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
