基于随机森林算法的阿尔茨海默症医学影像分类

Medical image classification for Alzheimer's disease diagnosis based on random forest algorithm

为实现阿尔茨海默症(AD)的医学影像分类,辅助医生对患者的病情进行准确判断,本研究对采集的34名AD患者、35名轻度认知障碍患者和35名正常对照组成员的功能磁共振影像进行特征提取和分类,具体思路包括:首先利用皮尔逊相关系数计算脑区之间的功能连接,然后采用随机森林算法对被试不同脑区之间的功能连接进行重要性度量及特征选择,最后使用支持向量机分类器进行分类,利用十倍交叉验证估算分类准确率。实验结果显示,随机森林算法可以对功能连接特征进行有效分析,同时得到AD发病过程的异常脑区,基于随机森林和SVM建立的分类模型对AD、轻度认知障碍的识别具有较好的效果,分类准确率可达90.68%,相关结论可以为AD的早期临床诊断提供客观参照。

For accurately classifying the medical images of Alzheimer's disease(AD)and assisting the doctors in making an accurate diagnosis of the patient's condition,a computer-aided diagnosis method is proposed based on random forest algorithm.The functional magnetic resonance imaging(fMRI)data of 34 AD patients,35 patients with mild cognitive impairment(MCI)and 35 normal controls are collected for feature extraction and classification.Firstly,the functional connections between different brain regions are calculated using Pearson correlation coefficient.Then the importance of the functional connections between different brain regions is assessed and important features are selected by random forest algorithm.Finally,support vector machine classifier is used for classification,and ten-fold cross-validation for estimating the classification accuracy.The experimental results show that random forest algorithm can be use to effectively analyze the functional connection characteristics and obtain the abnormal brain regions of AD pathogenesis.The classification model based on random forest and support vector machine has a good effect on the recognition of AD and MCI,with a classification accuracy of 90.68%.The related experimental results provide an objective reference for the early clinical diagnosis ofAD.