Design of a Deep Convolutional Neural Network based Model for Alzheimer’s disease Classification
Alzheimer's disease is perhaps the most prevalent type of dementia, which begins with memory loss and eventually leads to death. This is an irreversible condition that mostly affects the elderly. The most recent advancements in multimodal neuroimaging data allowed for the detection of disease in real time, which was a significant development in neuroscience. The greater level of correlation between the brain images, on the other hand, posed a significant threat in the diagnosis. Among the current studies, the Deep Learning methodology has shown excellent results in image classification. As a result, it is used to classify brain images into five categories: Late Mild Cognitive Impairment (LMCI), Mild Cognitive Impairment (MCL), Early Mild Cognitive Impairment (EMCL), Cognitively Normal (CN), and Alzheimer's disease (AD), ensuring very specific and accurate diagnosis. For the classification method, three pre-trained networks, AlexNet, ResNet-18, and GoogLe Net, were updated and trained for 3000 images using the transfer learning approach. The three networks have all been educated on the same collection of images from the ADNI database.