[Preprint] A MuST for Consistency Regularization in Semi-Supervised Medical Image Segmentation

A semi-supervised multi-scale training framework that applies consistency regularization across decoder layers to use unlabeled data for medical image segmentation.