A novel framework based on dual-level domain mixing is proposed. The proposed framework consists of three stages. First, two kinds of data mixing methods are proposed to reduce domain gap in both region-level and sample-level respectively. We can obtain two complementary domain-mixed teachers based on dual-level mixed data from holistic and partial views respectively..
Recommended citation: ShuaiJun Chen, Xu Jia, Jianzhong He and etal. (2021). "Semi-supervised Domain Adaptation based on Dual-level Domain Mixing for Semantic Segmentation." CVPR 2021.