Enhancing Alzheimer's disease Diagnosis with Augmented OASIS-l MRI Data: A Deep Convolutional Neural Network Approach

Alzheimer's disease (AD) is a neurodegenerative condition marked by ongoing deterioration of the brain, leading to memory impairment and the degeneration of brain cells. This disorder substantially hinders an individual's capacity to perform daily activities. Despite ongoing research, identifying the precise cause of AD remains a challenge, and effective treatment options are currently limited. Nevertheless, detecting the condition at an early stage could enhance the well-being of those grappling with Alzheimer's disease. In this investigation, we utilize data augmentation methods on magnetic resonance imaging (MRI) scans obtained from the open access series of imaging studies (OASIS)-t repository. To address the issue of data imbalance. Specifically, we aim to enhance the dataset's utility by augmenting the available images. To facilitate accurate classification, this work introduces an innovative to-layer deep convolutional neural network (DCNN). Training this DCNN involves utilizing two distinct sets of datasets. Significantly, our results demonstrate that the augmented dataset, devoid of class imbalance, attains a competitive level of accuracy in comparison to the non-augmented dataset, which inherently exhibits class imbalance. These outcomes emphasize the capacity of data-augmentation to enhance the effectiveness of AD classification models.