Despite the remarkable progress in semi-supervised medical image segmentation methods based on deep learning, their application to real-life clinical scenarios still faces considerable challenges. For ...