通过多层特征注意力提升自动驾驶目标检测器的运行时故障预测能力。
LFA: Layer Feature Attention for Run-Time Introspection of 2D Object Detectors in Automated Driving
- 用注意力机制融合骨干网络多层特征，捕捉不同抽象层级信息。
- 在KITTI和BDD100K上优于单层特征方法，错误预测准确率显著提升。
- 适合需要高可靠性的自动驾驶系统，可解释性强，便于故障分析。
可靠的物体检测对自动驾驶至关重要，但即使是最先进的检测器也会出错，可能危及安全。通过预测检测器故障来触发备用机制或提醒人工操作者，可实现更安全的部署。然而，现有方法仅依赖最后一层特征或手工设计的统计量，忽略了早期层中蕴含的丰富视觉抽象信息。本文提出层特征注意力（LFA），一种轻量级的运行时故障诊断方法，通过注意力机制学习融合多个骨干网络层的特征。核心思想是：检测错误在不同特征层次的表现形式不同——低层特征捕捉细粒度细节，有助于识别小尺寸或被遮挡物体；高层特征则编码语义信息，支持场景理解。LFA以端到端方式学习各层重要性权重，不仅提升了错误预测性能，还实现了对故障指示特征层级的可解释分析。在KITTI和BDD100K数据集上的大量实验表明，LFA在多种检测器架构下均达到当前最优的故障预测表现，显著优于仅使用单层特征的方法。
原文摘要 · Abstract (English)
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 or hand-crafted statistics, discarding valuable information from earlier layers that capture different levels of visual abstraction. We propose Layer Feature Attention (LFA), a lightweight introspection method that learns to aggregate features from multiple backbone layers through an attention mechanism. Our key insight is that detection errors manifest differently across feature hierarchies: low-level layers capture fine-grained details essential for detecting small or occluded objects, while high-level layers encode semantic information for scene understanding. LFA learns layer importance weights end-to-end, enabling both improved error prediction and interpretable analysis of which feature levels are most indicative of detector failures. Extensive experiments on KITTI and BDD100K demonstrate that LFA achieves state-of-the-art introspection performance, outperforming single-layer baselines across multiple detector architectures.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。