基于Mean Teacher的半监督左心房分割算法分析  

Analysis of Semi supervised Left Atrial Segmentation Algorithm Based on Mean Teacher

作　　者：宋鑫 孙鹏 苏云天 黄杰 陈真诚[1] SONG Xin;SUN Peng;SU Yuntian;HUANG Jie;CHEN Zhencheng(School of Electronic Engineering and Automation,Guilin University of Electronic Science and Technology,Guangxi 541004,China)

机构地区：[1]桂林电子科技大学电子工程与自动化学院,广西541004

出　　处：《电子技术（上海）》2024年第2期114-115,共2页Electronic Technology

基　　金：国家基金委区域创新联合基金重点项目(U22A2092);广西科技厅创新驱动发展专项项目(GuikeAA19254003)。

摘　　要：阐述基于Mean Teacher设计一种新的半监督分割框架,用于从3D MRI图像中分割出左心房。它使用改进的V-Net网络作为骨干网络,缓解原始模型计算时参数过多、计算量过大的状况,结合置信学习模块来解决噪声标签导致的分割性能下降问题。该方法相较其他半监督方法,具有更好的分割效果。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.

关 键 词：半监督学习 Mean Teacher模型 左心房分割 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TP391.41[自动化与计算机技术—控制科学与工程]