Semi-supervised boundary-aware medical image segmentation via symmetric boundary-foreground collaboration

Download PDF
Open Webpage

Xibin Jia, Wang Zhang, Luo Wang, Chuanxu Yang, Xunjie Yin, Hao Jia, Yiming Zheng, Zhenghan Yang, Dawei Yang, Min Hong, Hui Xu

Published: 01 Jan 2026, Last Modified: 12 Nov 2025Expert Syst. Appl. 2026EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•We realize effective medical image segmentation via boundary-foreground collaboration.•We propose cross-task and cross-model consistency regularization for unlabeled data.•Semi-supervision is built by supervised optimization and consistency regularization.•Extensive experiments on 5 datasets reveal the superiority of the proposed method.