Thyroid ultrasound diagnosis improvement via multi-view self-supervised learning and two-stage pre-training

Jian Wang, Xin Yang, Xiaohong Jia, Wufeng Xue, Rusi Chen, Yanlin Chen, Xi-Liang Zhu, Lian Liu, Yan Cao, Jianqiao Zhou, Dong Ni, Ning Gu

Abstract: Highlights•We proposed a multi-view SSL method that is not constrained by paired views.•We adopted a two-stage pre-training strategy on thyroid ultrasound images.•Extensive experiments were conducted on a large thyroid ultrasound image dataset.