VTSRM:一种基于SVM-RFE和MRMR的AD MRI医学图像分类方法  

VTSRM:an Method of AD MRI Medical Image Classification Based on SVM-RFE and MRMR

作　　者：周琼 陈梅[1,2] 李晖[1,2] 戴震宇[1,2] ZHOU Qiong;CHEN Mei;LI Hui;DAI Zhenyu(College of Computer Science and Technology,Guizhou University,Guiyang 550025;Guizhou Engineering Lab for ACMIS,Guizhou University,Guiyang 550025)

机构地区：[1]贵州大学计算机科学与技术学院,贵阳550025 [2]贵州大学贵州省先进计算与医疗信息服务工程实验室,贵阳550025

出　　处：《计算机与数字工程》2019年第6期1452-1458,共7页Computer & Digital Engineering

基　　金：国家自然科学基金项目“云平台上基于海量医学图像并行数据挖掘的计算机辅助诊断技术研究”(编号:61562010);贵州省自然科学基金项目(编号:20167427)资助

摘　　要：为了准确地识别阿尔兹海默症(Alzheimer’s Disease,AD),轻度认知障碍(Mild Cognitive Impairment,MCI)和正常个体(Normal Controls,NC),论文实现了一种基于SVM-RFE和MRMR的AD MRI医学图像分类方法 VTSRM。该方法首先提取出MRI医学图像的纹理特征和形态学特征,然后利用基于支持向量机递归特征消除算法(SVM-RFE)和最小冗余最大相关(MRMR)技术的特征选择算法SRM选择出最优特征子集,并使用SVM分类算法对AD,MCI,NC进行分类。美国公共阿尔茨海默病神经影像学数据集上的实验证明了论文方法的有效性。In order to accurately distinguish Alzheimer’s disease(AD),Mild Cognitive Impairment(MCI)and normal controls(NC),this paper implements a method VTSRM based on SVM-RFE and MRMR for AD MRI medical image classification.The method first extracts the texture features and morphological features of MRI medical images,and then selects the optimal feature subset by using the feature selection algorithm SRM which based on Support Vector Machine Recursive Feature Elimination Algorithm(SVM-RFE)and Minimum Redundant Maximum Correlation(MRMR). Last,SVM classification algorithm is used to classify AD,MCI,NC. Experiments on the neuroimaging dataset of public Alzheimer’s disease in the United States demonstrate the effectiveness of the proposed method.

关 键 词：MRI 形态学特征 纹理特征 SVM-RFE 最小冗余最大相关 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]