一种分割脑部磁共振图像的FCM改进算法  

AN IMPROVED FCM ALGORITHM FOR SEGMENTING BRAIN MRI

作　　者：杨阳 沈艳冰 李竹[1] Yang Yang;Shen Yanbing;Li Zhu(School of Physics and Information Engineering,Shanxi Normal University,Linfen 041004,Shanxi,China)

机构地区：[1]山西师范大学物理与信息工程学院,山西临汾041004

出　　处：《计算机应用与软件》2020年第3期231-235,共5页Computer Applications and Software

基　　金：国家自然科学基金青年科学基金项目(61503231)。

摘　　要：脑部MRI的快速准确分割是脑部疾病临床诊断过程的关键步骤之一。针对FCM算法部分参数设置影响分割结果和鲁棒性差的缺陷,提出一种基于非局部空间信息的快速模糊C均值核聚类改进算法,并应用于脑部MRI分割中。依次通过直方图、K-means算法、核函数、基于积分图的非局部空间信息解决了部分初始参数值难以控制、抗噪性差和运算效率低等问题。实验表明,该算法错分率低至2.0%,运行时间平均减少至13.89 s。The rapid and accurate segmentation of brain MRI is one of the key steps in the clinical diagnosis of cerebrovascular diseases.Aiming at the defects of FCM algorithm that some parameters affect the segmentation results and poor robustness,an improved fast fuzzy C-means kernel clustering algorithm based on non-local spatial information is proposed and applied to brain MRI segmentation.Through the histogram,K-means algorithm,kernel function and non-local spatial information based on integral graph,we sloved some defects:difficult control of initial parameter values,poor noise resistance and low operational efficiency.Experiments show that the proposed algorithm has a misclassification rate as low as 2.01%,and its running time is reduced to 13.89 s on average.

关 键 词：直方图 核函数 基于积分图的非局部空间信息 脑部磁共振图像分割 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]