A Diagnostic Classification Framework for Alzheimer's Disease: An Integrated Approach with ESIHE, SVM, and Wavelet Features

Alzheimer's disease (AD) is an irreversible neurodegenerative condition affecting the brain, resulting in cognitive decline, memory loss, and other neurological impairments, with no current cure or restorative treatment available. The primary objective of our study is to leverage machine learning (ML) methodologies for the precise and prompt identification of AD through medical images. The method utilizes structural magnetic resonance imaging (sMRI) scans derived from the Kaggle dataset, encompassing four distinct classes: NonDemented (ND), VeryMildDemented (VMD), MildDemented (MD), and ModerateDemented (MoD). Our methodology employs Exposure-based sub-Image histogram equalization (ESIHE) for preprocessing, enhancing image quality for subsequent analysis. The feature extraction process utilizes Discrete Wavelet Transform (DWT) to capture meaningful patterns from MR images. We proposed a classification based on Support Vector Machine (SVM) algorithm that involves two stages. In the first stage, three binary classifications were performed, while six binary classifications in the second stage. The automated classification achieved by our proposed model demonstrates compelling performance.