Pseudolabel guided pixels contrast for domain adaptive semantic segmentation

TL;DR

This paper introduces Pseudo-label Guided Pixel Contrast (PGPC), a novel contrastive learning framework for unsupervised domain adaptation in semantic segmentation that leverages pseudo-labels to improve feature diversity and accuracy.

Contribution

The paper proposes PGPC, a new contrastive learning method that addresses feature diversity issues in UDA for semantic segmentation, outperforming existing methods on standard benchmarks.

Findings

Achieves 5.1% and 4.6% relative mIoU improvements on GTA5 to Cityscapes and SYNTHIA to Cityscapes tasks.

Outperforms existing UDA methods without increasing model complexity.

Enhances other UDA approaches when integrated, demonstrating versatility.

Abstract

Semantic segmentation is essential for comprehending images, but the process necessitates a substantial amount of detailed annotations at the pixel level. Acquiring such annotations can be costly in the real-world. Unsupervised domain adaptation (UDA) for semantic segmentation is a technique that uses virtual data with labels to train a model and adapts it to real data without labels. Some recent works use contrastive learning, which is a powerful method for self-supervised learning, to help with this technique. However, these works do not take into account the diversity of features within each class when using contrastive learning, which leads to errors in class prediction. We analyze the limitations of these works and propose a novel framework called Pseudo-label Guided Pixel Contrast (PGPC), which overcomes the disadvantages of previous methods. We also investigate how to use more information from target images without adding noise from pseudo-labels. We test our method on two standard UDA benchmarks and show that it outperforms existing methods. Specifically, we achieve relative improvements of 5.1% mIoU and 4.6% mIoU on the Grand Theft Auto V (GTA5) to Cityscapes and SYNTHIA to Cityscapes tasks based on DAFormer, respectively. Furthermore, our approach can enhance the performance of other UDA approaches without increasing model complexity. Code is available at https://github.com/embar111/pgpc