链式窥视：面向视频理解的搜索引导渐进性对象基础推理
Chain-of-Glimpse: Search-Guided Progressive Object-Grounded Reasoning for Video Understanding
AI总结：
本文提出Chain-of-Glimpse框架，通过搜索引导的渐进推理解决视频中对象变化问题，提升多步骤决策的准确性和可解释性。
摘要：
视频理解需要在不同帧间识别和推理语义区分度高的视觉对象，但现有对象无关方法难以有效处理时间变化带来的显著对象变化。为此，我们引入Chain-of-Glimpse，一种搜索引导的渐进性对象基础推理框架，通过将每个推理步骤明确锚定到特定视觉证据区域，实现组合性和多步骤决策。形式上，Chain-of-Glimpse将视频推理视为逐步过程，逐步构建围绕任务相关视觉对象的空间基础轨迹，从而减少对显著性驱动线索的过度依赖。具体而言，Chain-of-Glimpse包含一个搜索引导的控制器，通过强化学习优化，以格式奖励显著激励基础能力，以迭代地基础视觉证据区域并形成可靠的推理轨迹，产生准确且可解释的多步骤决策。在域内NExTQA和域外Video-Holmes、CG-Bench Reasoning和VRBench基准测试中，广泛评估表明Chain-of-Glimpse在多样化视频推理任务中表现出一致的性能提升、鲁棒性和泛化能力。
英文摘要：
Video understanding requires identifying and reasoning over semantically discriminative visual objects across frames, yet existing object-agnostic solutions struggle to effectively handle substantial object variations over time. To address this, we introduce Chain-of-Glimpse, a search-guided progressive object-grounded reasoning framework that explicitly anchors each reasoning step to specific visual evidence regions, enabling compositional and multi-step decision-making. Formally, Chain-of-Glimpse formulates video reasoning as a step-by-step process that incrementally builds spatially grounded traces around task-relevant visual objects, thereby mitigating over-reliance on saliency-driven cues. Specifically, Chain-of-Glimpse features a search-guided controller, optimized via reinforcement learning with a format reward that significantly incentivizes grounding capability, to iteratively ground visual evidence regions and form reliable reasoning trajectories, yielding accurate and interpretable multi-step decisions. Extensive evaluations on both in domain NExTQA and out-of-domain Video-Holmes, CG-Bench Reasoning, and VRBench benchmarks demonstrate consistent performance gains, robustness and generalization of Chain-of-Glimpse across diverse video reasoning tasks.
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。
