[论文] CARA: Concept-Aware Risk Attention for Interpretable Collision Anticip...

论文概要

研究领域: CV 作者: Zhishan Tao, Ruoyu Wang, Yucheng Wu, Enjun Du, Yilei Yuan, Sherwin Ho, Yue Su, Jinbo Su, Yi Hong 发布时间: 2026-07-24 arXiv: 2607.22494

中文摘要

自动驾驶中的碰撞预判不仅需要准确的早期预警，还需要对正在跟踪哪些风险因素以及风险如何随时间演变进行可解释推理。现有方法在这方面存在不足：特征驱动模型不透明，事后解释通常缺乏保真度，而基于概念的方法大多针对静态识别设计而非动态驾驶场景。本文提出CARA（概念感知风险注意力），一种用于碰撞预判的内在可解释时空框架。CARA从事故叙述中派生领域基础的风险概念，通过视觉-语言相似性将它们与视频帧对齐，并将它们组织为演变的概念轨迹。这些轨迹提供显式的风险证据，指导空间注意力、时间注意力和预判，允许语义概念直接影响模型关注何处以及如何随时间预测风险。通过将语义风险因素视为动态中间证据而非辅助事后解释，CARA将可解释性与预测过程紧密耦合。在三个基准测试上的大量实验表明，CARA持续提高了预判准确性和预警提前量，优于强基线，同时提供稀疏且语义基础的概念证据。

原文摘要

Collision anticipation in autonomous driving requires not only accurate early warnings but also interpretable reasoning about what risk factors are being tracked and how risk evolves over time. Existing methods fall short in this regard: feature-driven models are opaque, post-hoc explanations often lack fidelity, and concept-based methods are mostly designed for static recognition rather than dynamic driving scenes. We propose CARA (Concept-Aware Risk Attention), an intrinsically interpretable spatio-temporal framework for collision anticipation. CARA derives domain-grounded risk concepts from accident narratives, aligns them with video frames via vision-language similarity, and organizes them into evolving concept trajectories. These trajectories provide explicit risk evidence that guides...

--- *自动采集于 2026-07-28*

#论文 #arXiv #CV #小凯