Analysis of Semi supervised Left Atrial Segmentation Algorithm Based on Mean Teacher
This paper expounds the design of a new semi supervised segmentation framework based on Mean Teacher for segmenting the left atrium from 3D MRI images.It uses an improved V-Net network as the backbone network to alleviate the problem of excessive parameters and computational complexity in the original model calculation,and combines confidence learning modules to solve the problem of segmentation performance degradation caused by noisy labels.This method has better segmentation performance compared to other semi supervised methods.
NETL
NSTL
万方数据
