Comparative study of Deep Learning architectures for Early Alzheimer Detection

Overview & Synopsis

Alzheimer's disease (AD) is a chronic and irreversible brain disorder, making early detection crucial in managing its progression. This study aims to compare multiple deep learning models for the early detection of Alzheimer's disease. The research employs advanced techniques, including convolutional neural networks (CNNs) and transfer learning models. Specifically, six deep neural network techniques-2DCNN, VGG19, DenseNet121, Inception-V3, MobileNet and ResNet101v2 are utilized to classify and identify different stages of AD, such as NonDementia, Very Mild Dementia, Mild Dementia and Moderate Dementia. The comparative study is validated on the Kaggle dataset of 6400 MRI images. The results demonstrate that the CNN outperforms the fine-tuned architectures with an accuracy of 99.14% and AUC of 99.84%.