Safety Metrics for Semantic Segmentation in Autonomous Driving

TL;DR

This paper introduces safety-aware correctness and robustness metrics for semantic segmentation in autonomous driving, moving beyond pixel-level measures to account for the safety criticality of pixel clusters and locations.

Contribution

It proposes novel safety metrics for semantic segmentation that consider pixel clustering and location, enhancing safety assessment in autonomous driving systems.

Findings

Metrics reflect safety criticality based on pixel clustering and location.

Validated on autonomous driving dataset, demonstrating practicality.

Enhances existing safety evaluation methods for semantic segmentation.

Abstract

Within the context of autonomous driving, safety-related metrics for deep neural networks have been widely studied for image classification and object detection. In this paper, we further consider safety-aware correctness and robustness metrics specialized for semantic segmentation. The novelty of our proposal is to move beyond pixel-level metrics: Given two images with each having N pixels being class-flipped, the designed metrics should, depending on the clustering of pixels being class-flipped or the location of occurrence, reflect a different level of safety criticality. The result evaluated on an autonomous driving dataset demonstrates the validity and practicality of our proposed methodology.