Synthetic outlier generation for anomaly detection in autonomous driving

TL;DR

This paper improves anomaly detection in autonomous driving by modifying training strategies for semantic segmentation models, resulting in enhanced detection performance and a simplified yet effective detector.

Contribution

It introduces novel training modifications to the DenseHybrid model and proposes a simplified detector that outperform previous methods.

Findings

Significant performance improvements in anomaly detection.

A simplified detector achieves comparable results to complex models.

Enhanced safety in autonomous driving through better anomaly detection.

Abstract

Anomaly detection, or outlier detection, is a crucial task in various domains to identify instances that significantly deviate from established patterns or the majority of data. In the context of autonomous driving, the identification of anomalies is particularly important to prevent safety-critical incidents, as deep learning models often exhibit overconfidence in anomalous or outlier samples. In this study, we explore different strategies for training an image semantic segmentation model with an anomaly detection module. By introducing modifications to the training stage of the state-of-the-art DenseHybrid model, we achieve significant performance improvements in anomaly detection. Moreover, we propose a simplified detector that achieves comparable results to our modified DenseHybrid approach, while also surpassing the performance of the original DenseHybrid model. These findings demonstrate the efficacy of our proposed strategies for enhancing anomaly detection in the context of autonomous driving.