Abstract: Semi-supervised learning (SSL) addresses the scarcity of annotated data in medical image segmentation by leveraging unlabeled samples to enhance model training. Currently, some methods ...