The Classification of Alzheimer’s Disease: A Transfer Learning Approach

Manual diagnosis of Alzheimer’s Disease (AD) from medical images is challeng-ing and prone to human subjectivity and errors. This paper explores the application of transfer learning, utilizing pre-trained convolutional neural networks, namely, VGG16 to extract the features from magnetic resonance imaging images of four stages of AD. The ability of four machine learning classifiers, i.e., Support Vector Machine (SVM), Logistic Regression (LR) and k-Nearest Neighbours (kNN) to class the different stages based on the extracted features are investigated. It was demonstrated that the VGG16+LR pipeline holds a promising ability in discerning the classes. The preliminary results suggest that the proposed approach is a suitable method for computer-aided diagnosis.