Alzheimer Detection Using MRI Imaging Modality

Abstract

Alzheimer’s disease (AD) is a neurological disease that affects memory and livelihood of the people that are diagnosed with it. Many different imaging modalities have been used to help diagnose the disease. Each of these modalities offers something different towards the detection and possible treatments for AD. In this project, we developed a new approach based on mathematical and image processing techniques for better classification of AD. We proposed to reduce the high dimensional magnetic resonance imaging (MRI) image vector space to 150 dimensions using principal component analysis (PCA). In order to categorize the reduced dimensions from PCA for progression of AD, we employed a multiclass neural network. The neural network was trained initially on 38 diagnosed MRIs obtained from OASIS MRI database. We then tested our trained neural network on the entire set of 86 MRIs provided by OASIS dataset to confirm the accuracy of diagnosis by our system. Our results produced nearly 92% accuracy in AD diagnosis and classification