Design of A Convolutional Neural Network System to Increase Diagnostic Efficiency of Alzheimer's Disease

摘要

The most common degenerative neural disease,Alzheimer's disease(AD),is insidious and almost always requires imaging modalities to be diagnosed early.MRI is the most common one used,but requires timely interpretation.Here we develop a convolutional neural network(CNN)-based system that determines whether a brain MR image has AD or normal.First,feature extraction is performed to separate various parts of the brain.Then,the data is processed to differentiate normal brain from AD brain,solely using MR image.Finally,the neural network is supplemented using data from the patient's history and physical examination.In this first phase,we were able to extract features from the brain MR image,initially by masking the image and separating the white matter,grey matter,and cerebrospinal fluid called the grey level cooccurrence method(GLCM).This method is able to using a convolutional neural network.