研究人员开发了一种使用自我中心视觉和视觉语言模型（VLMs）预测行人过马路意图的新方法。通过将任务构建为视觉问答，他们对 VLMs 进行了微调，使其性能显著优于现有的基于 Transformer 的模型。包含的眼动和自我运动等上下文线索进一步提高了预测准确性，为这一安全关键应用树立了新的最先进水平。 AI
影响 为行人意图预测树立了新的最先进水平，有可能提高自动驾驶安全系统。
排序理由 该集群包含一篇详细介绍新研究方法和基准测试结果的学术论文。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2606.09142v1 Announce Type: cross Abstract: Egocentric vision offers a first-person view of human perception and decision making, yet its potential for traffic-safety prediction remains underexplored. In this work, we study the decoding of pedestrian crossing intentions fro…
Egocentric vision offers a first-person view of human perception and decision making, yet its potential for traffic-safety prediction remains underexplored. In this work, we study the decoding of pedestrian crossing intentions from short egocentric video clips. We approach this b…
Egocentric vision offers a first-person view of human perception and decision making, yet its potential for traffic-safety prediction remains underexplored. In this work, we study the decoding of pedestrian crossing intentions from short egocentric video clips. We approach this b…