Research on Unsupervised Domain Adaptation to Reduce Labelling Costs

点击翻译按钮获取中文摘要

During WACV 2023, we introduced a classifier-based backprop-induced weighting of the feature space. This approach has two main advantages. Firstly, it lets the domain classifier focus on features that are important for the classification, and, secondly, it couples the classification and adversarial branch more closely. Furthermore, we introduce an iterative label distribution alignment method, that employs results of previous runs to approximate a class-balanced dataloader.We also are currently finalizing a paper based on domain adaptation via incremental domain enlargement, that incrementally employs more and more target domain specific features.Further two papers in collaboration with other researchers have been accepted for ICPR 2022.