Distribution-Constrained Semi-Supervised Learning for Medical Organ Segmentation
A Distribu-tion-Constrained Semi-Supervised Segmentation Model (DCSSM) is introduced to capture positional correlations in medical images and achieves the best performance across multiple segmentation models on 5 medical segmentation datasets.
Linjiang Liu, Shu-Ting Huang, Bojian Yu
· 0 citations