Detection of Alzheimer’s disease in MRI images using different transfer learning models and improving the classification accuracy

Abstract

Alzheimer's disease (AD) is a neurodegenerative illness that damages brain cells and impairs a patient's memory over time. If detected early so, the patient can avoid permanent memory loss and further damage to their brain cells. Various automated technologies and techniques have been developed in recent years for the detection of Alzheimer's disease (AD). Methods that focus on rapid, accurate, and early identification of the condition in order to reduce the negative impact on a patient's mental health are available. Medical imaging systems for Alzheimer’s disease (AD) diagnostic performance have been greatly improved by machine learning models. There is a major difficulty with multi-class classification, however, which is the presence of brain structural characteristics that are extremely closely associated. It is possible to improve deep learning by increasing the number of layers and including features and classifiers at all levels of the classification hierarchy. Nevertheless, the vast majority of deep learning models (like traditional CNN model) fail to deliver acceptable results in real-world situations.