基于内容的医学图像检索综述  

Review of Content-Based Medical Image Retrieval

作　　者：刘桂慧 Liu Guihui(North Minzu University,Yinchuan Ningxia 750021,China)

机构地区：[1]北方民族大学,宁夏银川750021

出　　处：《信息与电脑》2020年第15期51-53,共3页Information & Computer

基　　金：北方民族大学研究生创新项目(项目编号:YCX19069)。

摘　　要：随着医学成像技术的不断发展,患者的各类扫描图像呈爆发式增长,医务人员的工作量日益繁重,基于内容的医学图像检索(Content-based Medical Image Retrieval,CBMIR)系统可以精准地辅助医务工作者决策,减轻医务工作者的负担。基于此,笔者综述了CBMIR方法,着重探讨了特征提取、特征融合、相似性度量等的主流算法,跟踪其研究成果,并对CBMIR技术的未来研究方向进行了展望。With the continuous development of medical imaging technology,various types of scanned images of patients have exploded,and the workload of medical staff has become increasingly heavy.Content-based Medical Image Retrieval (CBMIR) technology can assist medical workers to make decisions more accurately and reduce the burden.This article reviews the content-based medical image retrieval methods,focuses on the relevant mainstream algorithms of feature extraction,feature fusion,similarity measurement and other core technologies,tracks their research results,and prospects the future research direction of CBMIR technology.

关 键 词：医学图像检索 特征提取 特征融合 相似性度量 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP183[自动化与计算机技术—计算机科学与技术]