Abstract

Info

Author(s):: ZHOU Wen; WANG Yu; XIAO Hongbing; CAO Lihong; Key Laboratory of Food Safety Big Data Technology, School of Computer and Information Engineering, Beijing Technology and Business University, Beijing 100048, China

Keywords:: ; Alzheimer’s disease; structural magnetic resonance imaging; kernel principal component analysis; feature extraction; machine learning

Abstract:: Alzheimer’s disease (AD) is a degenerative disease of nervous system development with insidious onset. Using magnetic resonance imaging (MRI) and computer technology for the assisted diagnosis of AD patients is a new topic which is gradually explored at present. Preprocessing and correlation analysis on MRI image are firstly performed. Then kernel principal component analysis (KPCA) is used to extract features of brain gray matter images. Then the extracted features are classified with Adaboost algorithm, and the result was compared with that classified by principal component analysis. Experimental results in AD Neuroimaging Initiative Database which contains brain structural MRI images of 116 AD patients, 116 patients with mild cognitive impairment, and 117 normal controls show that machine learning can effectively assist the diagnosis of brain diseases in AD patient. Compared with principal component analysis, KPCA algorithm is more complete for the feature extraction and more accurate for the classification. Using the proposed method in AD diagnosis can obtain more robust assisted diagnosis results.