Leveraging labelled data knowledge: A cooperative rectification learning network for semi-supervised 3D medical image segmentation
Abstract: Highlights•Innovative semi-supervised learning for3D medical image segmentation.•New multi-prototype class prior to rectify pseudo-labels at the voxel level.•Novel prototype and image feature interactions to improve pseudo-label accuracy.•Uncertain features are optimised by aligning with unassertive class representations.•Experiments from three different datasets demonstrate the superiority of our method.
Loading