Avoiding Negative Transfer for Semantic Segmentation of Remote Sensing Images

Reducing the feature distribution shift caused by the factor of visual-environment changes, named visual-environment changes (VE-changes), is a hot issue in domain adaptation learning. However, in the semantic segmentation task of remote sensing imageries, besides VE-changes, the change of semantic-scene changes (SS-changes) is another factor raising the domain gap, which brings the label distribution shift. For example, although urban and rural share the same landcover label, there is still a gap in label distribution. If there is little relation that can be found in neither feature nor label space, forcibly adapting to a new domain could have a high risk of negative transfer. Hence, we propose a new Transitive Domain Adaptation method for Remote Sensing (TDARS) images. First, we introduce an intermediate domain to enlarge the relation between the given source and target domains. Second, we learn from primary and nonprimary confident classes to increase the likelihood of transferring valuable information. As a result, TDARS enables the given source and target domains to be connected through the selected intermediate domain and performs effective knowledge transfer among all domains. The proposed method is evaluated on three domain adaptation datasets of remote sensing images. Extensive experiments show that the approach can effectively handle the domain shift problem from remote sensing images compared to other state-of-the-art domain adaptation methods.