Efficient Model based on Deep Learning for the Classification of Dementia

Abstract

Alzheimer's disease to severe dementia are all forms of dementia, which are neurodegenerative disorders that affect brain memory. The most prevalent type of dementia that impairs thinking and memory is Alzheimer's disease. Network width, depth, and resolution are scaled uniformly by Efficient Net using a set of compound coefficients. This study uses callback functions to diagnose four types of dementia in order to slow down the learning process and increase accuracy. Additionally, contrast CNN and the Inception V3 model with the EfficientNet deep learning model. Using EfficienNetB2, classification accuracy is approximately 99.3%.