Shallow Features Guide Unsupervised Domain Adaptation for Semantic Segmentation at Class Boundaries

TL;DR

This paper introduces a low-level feature adaptation method combined with boundary-aware data augmentation to improve the sharpness and accuracy of semantic segmentation masks in unsupervised domain adaptation, especially at class boundaries.

Contribution

It proposes a novel low-level adaptation strategy and boundary-focused data augmentation that enhance segmentation quality in domain adaptation tasks, particularly at class boundaries.

Findings

Improved segmentation accuracy along class boundaries.

Effective integration with existing adaptation frameworks.

Significant performance gains demonstrated through experiments.

Abstract

Although deep neural networks have achieved remarkable results for the task of semantic segmentation, they usually fail to generalize towards new domains, especially when performing synthetic-to-real adaptation. Such domain shift is particularly noticeable along class boundaries, invalidating one of the main goals of semantic segmentation that consists in obtaining sharp segmentation masks. In this work, we specifically address this core problem in the context of Unsupervised Domain Adaptation and present a novel low-level adaptation strategy that allows us to obtain sharp predictions. Moreover, inspired by recent self-training techniques, we introduce an effective data augmentation that alleviates the noise typically present at semantic boundaries when employing pseudo-labels for self-training. Our contributions can be easily integrated into other popular adaptation frameworks, and extensive experiments show that they effectively improve performance along class boundaries.