Add to Chrome
Add to Firefox
Add to EdgeGet away from Style: Category-Guided Domain Adaptation for Semantic Segmentation
Mar 29, 2021
Unsupervised domain adaptation (UDA) becomes more and more popular in tackling real-world problems without ground truth of the target domain. Though a mass of tedious annotation work is not needed, UDA unavoidably faces the problem how to narrow the domain discrepancy to boost the transferring performance. In this paper, we focus on UDA for semantic segmentation task. Firstly, we propose a style-independent content feature extraction mechanism to keep the style information of extracted features in the similar space, since the style information plays a extremely slight role for semantic segmentation compared with the content part. Secondly, to keep the balance of pseudo labels on each category, we propose a category-guided threshold mechanism to choose category-wise pseudo labels for self-supervised learning. The experiments are conducted using GTA5 as the source domain, Cityscapes as the target domain. The results show that our model outperforms the state-of-the-arts with a noticeable gain on cross-domain adaptation tasks.
