Triple-task mutual consistency for semi-supervised 3D medical image segmentation

Yantao Chen, Yong Ma, Xiaoguang Mei, Lin Zhang, Zhigang Fu, Jiayi Ma

Abstract: Highlights•We propose a triple-task mutual consistency framework for semi-supervised segmentation.•A novel regression task is proposed to enhance the edge awareness of the model.•We overcome the task dependence and gradient propagation problems in previous studies.•Our method achieves superior performance on two public medical datasets.