学习纹理不变的表示以进行语义分割的领域自适应

Mar, 2020

学习纹理不变的表示以进行语义分割的领域自适应

Learning Texture Invariant Representation for Domain Adaptation of Semantic Segmentation

Myeongjin Kim, Hyeran Byun

TL;DR本文提出了一种使用风格迁移算法来多样化合成图像纹理，然后通过自训练来微调模型以适应目标域纹理的领域自适应方法，以实现更好的语义分割性能。

Abstract

Since annotating pixel-level labels for semantic segmentation is laborious, leveraging synthetic data is an attractive solution. However, due to the domain gap between synthetic domain and real domain, it is chal