How to Relieve Distribution Shifts in Semantic Segmentation for Off-Road Environments

2026-05-28

AI Summary

This work addresses the challenge of distribution shift in semantic segmentation caused by domain discrepancies and sensor perturbations in off-road environments. To this end, the authors propose ST-Seg, a novel framework that explicitly expands the source domain distribution by enhancing its diversity through style augmentation. The method further introduces texture regularization to stabilize local texture representations and leverages a deep texture manifold to improve generalization to unseen target domains. Experimental results demonstrate that ST-Seg significantly outperforms existing approaches across multiple off-road distribution shift scenarios, achieving substantial gains in both segmentation robustness and real-world navigation performance.

Abstract

Semantic segmentation is crucial for autonomous navigation in off-road environments, enabling precise classification of surroundings to identify traversable regions. However, distinctive factors inherent to off-road conditions, such as source-target domain discrepancies and sensor corruption from rough terrain, can result in distribution shifts that alter the data differently from the trained conditions. This often leads to inaccurate semantic label predictions and subsequent failures in navigation tasks. To address this, we propose ST-Seg, a novel framework that expands the source distribution through style expansion (SE) and texture regularization (TR). Unlike prior methods that implicitly apply generalization within a fixed source distribution, ST-Seg offers an intuitive approach for distribution shift. Specifically, SE broadens domain coverage by generating diverse realistic styles, augmenting the limited style information of the source domain. TR stabilizes local texture representation affected by style-augmented learning through a deep texture manifold. Experiments across various distribution-shifted target domains demonstrate the effectiveness of ST-Seg, with substantial improvements over existing methods. These results highlight the robustness of ST-Seg, enhancing the real-world applicability of semantic segmentation for off-road navigation.

Problem

• distribution shift
• semantic segmentation
• off-road environments
• domain discrepancy
• sensor corruption

Innovation

• style expansion
• texture regularization
• distribution shift
• semantic segmentation
• off-road navigation