BriefGPT.xyz
Apr, 2022
语义分割的无监督对比域自适应
Unsupervised Contrastive Domain Adaptation for Semantic Segmentation
HTML
PDF
Feihu Zhang, Vladlen Koltun, Philip Torr, René Ranftl, Stephan R. Richter
TL;DR
通过对比学习实现特征对齐,提出了一种跨领域适应的标签扩展方法,结果表明,我们的方法在Cityscapes数据集上的表现始终优于最先进的方法。
Abstract
semantic segmentation
models struggle to generalize in the presence of domain shift. In this paper, we introduce
contrastive learning
for
feature
→