Alzheimer's disease classification: a comprehensive study.

Alzheimer's disease (AD) is a neurodegenerative disease that is well-known for causing continuous loss of memory, cognition, and other higher brain functions. AD is not a single disease, but rather a group of related diseases with similar characteristics. The use of deep neural network-based pattern classification techniques, such as convolutional neural networks, is effective in classifying patients into different sub-types of AD and in distinguishing the different stages of severity of the disease. in the medical field, early detection of its start can be quite beneficial. This article focuses on the early detection of various stages of cognitive aging and AD using neuroimaging and transfer learning (TL). Images of imagery via resonance magnetic (IRM) obtained from a Kaggle database called Alzheimer's Dataset (4 class of Images) with several classes of non-dementia (NONDEM), very mild dementia(VERDEM), mild dementia(MILDEM), moderate dementia(MODDEM) are classified using a transfer learning approach. In this work, we compare the classification performance of six pre-trained networks, which are VGG-19, VGG-16, ResNet-50, InceptionV3, Xception, and DenseNet169. They were enthralled and tested using 6400 images from the Kaggle data pool. The confusion matrix and its parameters are used to assess the classification performance of these six networks. VGG-19, VGG-16, Inception-V3, Xception, ResNet-50, and DenseNet169 all have 92.86%, 92.83%, 91.04%, 90.57%, 85.99%, and 88.64% overall precision in MA detection, respectively. [ABSTRACT FROM AUTHOR]