Multi-rater Prism:Learning self-calibrated medical image segmentation from multiple raters  

多重标注梭镜:从多个标注中学习自校准临床影像图像分制

作　　者：Junde Wu Huihui Fang Jiayuan Zhu Yu Zhang Xiang Li Yuanpei Liu Huiying Liu Yueming Jin Weimin Huang Qi Liu Cen Chen Yanfei Liu Lixin Duan Yanwu Xu Li Xiao Weihua Yang Yue Liu 吴俊德;方慧卉;朱嘉源;张雨;李响;刘元培;刘慧颖;金玥明;黄为民;刘琦;陈岑;刘艳飞;段立新;许言午;肖力;杨卫华;刘玥

机构地区：[1]School of Future Technology,South China University of Technology,Guangzhou 511442,China [2]Pazhou Lab,Guangzhou 510320,China [3]The University of Oxford,Oxford OX14AL,UK [4]Cardiovascular Disease Center,Xiyuan Hospital of China Academy of Chinese Medical Sciences,Beijing 100091,China [5]State Key Laboratory of Pulsed Power Laser Technology,College of Electronic Engineering,National University of Defense Technology,Hefei 230037,China [6]Shenzhen Institute for Advanced Study,University of Electronic Science and Technology of China,Shenzhen 518110,China [7]The University of Hong Kong,Hong Kong 999077,China [8]Institute for Infocomm Research,A*STAR,Singapore 138632,Singapore [9]National University of Singapore,Singapore 119276,Singapore [10]Sichuan Provincial People’s Hospital,University of Electronic Science and Technology of China,Chengdu 611731,China [11]Shenzhen Eye Hospital,Jinan University,Shenzhen 518040,China

出　　处：《Science Bulletin》2024年第18期2906-2919,共14页科学通报（英文版）

基　　金：supported by the Excellent Young Science and Technology Talent Cultivation Special Project of China Academy of Chinese Medical Sciences(CI2023D006);the National Natural Science Foundation of China(82121003 and 82022076);Beijing Natural Science Foundation(2190023);Shenzhen Fundamental Research Program(JCYJ20220818103207015);Guangdong Provincial Key Laboratory of Human Digital Twin(2022B1212010004)。

摘　　要：In medical image segmentation,it is often necessary to collect opinions from multiple experts to make the final decision.This clinical routine helps to mitigate individual bias.However,when data is annotated by multiple experts,standard deep learning models are often not applicable.In this paper,we propose a novel neural network framework called Multi-rater Prism(MrPrism)to learn medical image segmentation from multiple labels.Inspired by iterative half-quadratic optimization,MrPrism combines the task of assigning multi-rater confidences and calibrated segmentation in a recurrent manner.During this process,MrPrism learns inter-observer variability while taking into account the image's semantic properties and finally converges to a self-calibrated segmentation result reflecting inter-observer agreement.Specifically,we propose Converging Prism(ConP)and Diverging Prism(DivP)to iteratively process the two tasks.ConP learns calibrated segmentation based on multi-rater confidence maps estimated by DivP,and DivP generates multi-rater confidence maps based on segmentation masks estimated by ConP.Experimental results show that the two tasks can mutually improve each other through this recurrent process.The final converged segmentation result of MrPrism outperforms state-of-the-art(SOTA)methods for a wide range of medical image segmentation tasks.The code is available at https://github.-com/WuJunde/MrPrism.

关 键 词：Medical image segmentation Multiple raters SELF-CALIBRATION Half-quadratic algorithm 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] R318[自动化与计算机技术—计算机科学与技术]