Alzheimer’s Disease Detection Using Resnet

Alzheimer’s disease is incurable. Early Alzheimer’s diagnosis helps with treatment and brain tissue preservation. Statistics and machine learning methods have been applied to diagnose AD. Clinical research uses MRI to diagnose AD. Advanced deep learning approaches have recently proven equivalent or even superior human-level performance. In addition to that, the strong computing power available today has had a huge impact on the algorithms used in deep learning, including medical picture processing. This paper aims to provide improved model performance for the early-stage diagnosis of AD. Using brain MRI data processing, we propose a deep CNN ResnetAD model for AD diagnosis. The result shows the best model accuracy of 90%, a training loss of 0.3924, a K-fold accuracy of 0.579, and a validation loss of 40%.