Transfer learning approach for Alzheimer’s disease diagnosis using MRI images

Abstract

Alzheimer's disease is the most prevalent type of dementia and is defined as a slow-progressing neurological disorder. As a first step, early diagnosis of Alzheimer's disease is crucial, then classification is required as a second step for patients to be offered the most effective treatment available. For testing and analyzing this research, the Alzheimer's Disease Neuroimaging Initiative (ADNI) Baseline dataset is used. In this study, we suggested utilizing a convolutional neural network (CNN) algorithm to diagnose Alzheimer's disease from MRI images using a supervised deep learning approach based on transfer learning. The implemented system examines two different CNN architectures, including VGG-16 and MobileNet-V2. According to our results, this study achieved the highest accuracy and F1-score with 99.71% and 100%, respectively.