DecoupleNet: Decoupled Network for Domain Adaptive Semantic Segmentation

TL;DR

DecoupleNet introduces a novel approach for unsupervised domain adaptation in semantic segmentation by decoupling feature alignment from segmentation focus, utilizing self-discrimination and online self-training to improve target domain performance.

Contribution

The paper proposes DecoupleNet, which reduces source overfitting and enhances segmentation focus, along with Self-Discrimination and Online Enhanced Self-Training for better target domain adaptation.

Findings

Outperforms existing state-of-the-art methods.

Effective in learning discriminative target features.

Components verified through extensive ablation studies.

Abstract

Unsupervised domain adaptation in semantic segmentation has been raised to alleviate the reliance on expensive pixel-wise annotations. It leverages a labeled source domain dataset as well as unlabeled target domain images to learn a segmentation network. In this paper, we observe two main issues of the existing domain-invariant learning framework. (1) Being distracted by the feature distribution alignment, the network cannot focus on the segmentation task. (2) Fitting source domain data well would compromise the target domain performance. To address these issues, we propose DecoupleNet that alleviates source domain overfitting and enables the final model to focus more on the segmentation task. Furthermore, we put forward Self-Discrimination (SD) and introduce an auxiliary classifier to learn more discriminative target domain features with pseudo labels. Finally, we propose Online Enhanced Self-Training (OEST) to contextually enhance the quality of pseudo labels in an online manner. Experiments show our method outperforms existing state-of-the-art methods, and extensive ablation studies verify the effectiveness of each component. Code is available at https://github.com/dvlab-research/DecoupleNet.