Domain generalization for mammographic image analysis with contrastive learning

Zheren Li, Zhiming Cui, Lichi Zhang, Sheng Wang, Chenjin Lei, Xi Ouyang, Dongdong Chen, Xiangyu Zhao, Chunling Liu, Zaiyi Liu, Yajia Gu, Dinggang Shen, Jie-Zhi Cheng

Abstract: Highlights•Deep learning models in mammography face limitations due to vendor style variations.•Our approach using contrastive learning to enhance model generalizability across diverse domains.•The multi-style and multi-view unsupervised self-learning scheme is carried out to seek robust feature embedding.•Comprehensive evaluation demonstrates significant improvements in the generalizability for critical mammographic tasks.