EfficientAD: A Deep Learning Approach for Multi-Stage AD Classification Using Transfer Learning

Alzheimer's disease (AD) is a neurodegenerative disease that has affected millions of elderly people worldwide. It is the most common cause of dementia among the aged population. It affects the cognitive skills of a person making it difficult to perform daily chores. Due to the incurability and progressive nature of AD, it has garnered a huge interest from the researchers. Neurologists treating AD patients analyze various neuroimaging scans using Computer-Aided Diagnosis (CAD) systems. These systems use DL algorithms to detect AD early and accurately. In this paper, various pretrained CNN models namely EfficientNetB7, VGG16, VGG19 and VGG Ensemble models were employed for binary and multistage classification of AD by using transfer learning. This paper achieved an accuracy of 97.88
^%
using the EfficientNetB7, 91.36%, 92.81 % and 95.16% using VGG16, VGG19 and VGG Ensemble models respectively.