LFA: Layer Feature Attention for Run-Time Introspection of 2D Object Detectors in Automated Driving 事件

LFA: Layer Feature Attention for Run-Time Introspection of 2D Object Detectors in Automated Driving arXiv:2606.00372v1 Announce Type: new Abstract: Reliable object detection is critical for automated driving, yet even state-of-the-art detectors inevitably make errors that can compromise safety. Introspection methods that predict detector failures enable safer deployment by triggering fallback mechanisms or alerting human operators. However, existing approaches rely solely on last-layer features