Abstract: Semi-supervised medical image segmentation aims to leverage a limited set of labeled images alongside a substantial volume of unlabeled images to train semantic segmentation models. Existing ...