Work

Quadruple-Consistency Vision Transformer for Medical Image Segmentation with Limited Number of Sparse Annotations