Robust self-supervised learning for source-free domain adaptation

• Published in: Signal, image and video processing Vol. 17; no. 5; pp. 2405 - 2413
• Main Authors: Tian, Liang, Zhou, Lihua, Zhang, Hao, Wang, Zhenbin, Ye, Mao
• Format: Journal Article
• Language: English
• Published: London Springer London 01.07.2023; Springer Nature B.V
• Subjects: Adaptation; Computer Imaging; Computer Science; Domains; Image Processing and Computer Vision; Labels; Masks; Multimedia Information Systems; Original Paper; Pattern Recognition and Graphics; Robustness; Self-supervised learning; Signal,Image and Speech Processing; Vision; Local structure clustering; Source-free domain adaptation; Self-supervised learning
• ISSN: 1863-1703; 1863-1711
• DOI: 10.1007/s11760-022-02457-z

Source-free domain adaptation (SFDA) is from unsupervised domain adaptation (UDA) and do apply to the special situation in reality that the source domain data is not accessible. In this subject, self-supervised learning is widely used in previous works. However, inaccurate pseudo-labels are hardly avoidable and that degenerates the adapted target model. In this work, we propose an effective method, named RS2L (robust self-supervised learning), to reduce the negative impact due to inaccurate pseudo-labels. Two strategies are adapted. The first is called structure-preserved pseudo-labeling strategy which generates much better pseudo-labels by stored predictions of k -closest neighbors. Another is self-supervised learning with mask. We use threshold masks to select samples for different operations, i.e., self-supervised learning and structure-preserved learning. For different masks, the threshold values are different. So it is not excluded that some samples participate in both two operations. Experiments on three benchmark datasets show that our method achieves the state-of-the-art results.