Jabli Mohamed Amine*, Moussa Mourad and Douik Ali
Alzheimer's disease is a common form of dementia that is often deadly, particularly among individuals over the age of 65. Early detection of Alzheimer's disease can improve patient outcomes, and machine learning techniques applied to Magnetic Resonance Imaging (MRI) scans have been utilized to aid in diagnosis and assist physicians. However, traditional machine learning approaches require the manual extraction of features from MRI images, a process that can be complicated and require expert input. To address this issue, we propose the use of a pre-trained Convolutional Neural Network (CNN) model, ResNet50, as a method of automatic feature extraction for the diagnosis of Alzheimer's disease using MRI images. We compare the performance of this model to conventional Softmax, Support Vector Machine (SVM), and Random Forest (RF) methods, evaluating the results using various metric measures such as accuracy. Our model outperformed other stateof- the-art models, achieving an accuracy range of 85.7% to 99% when tested with the ADNI MRI dataset.
Published Date: 2025-06-13; Received Date: 2024-07-07
