Multiclass-classification of Alzheimers Disease using 3-D CNN and Hyper-Parameter Optimization of Machine Learning Algorithms

Chinmay Pathak, Abhay Jidge, Vishal Mourya, Omkaresh Kulkarni, Bharati Dixit

Abstract: Dementia and its forms like Mild Cognitive Impairment and Alzheimer’s Disease are posing great defiance to society claiming many lives. Alzheimer’s Disease is the fifth leading cause of death in the world. The most frequently observed form of this dementia is Alzheimer’s Disease. Diagnosis of this disease is a very tedious task for the doctors, which can lead to errors in the judgment. This paper proposes multi-class classification by using MRI Images of subjects into 3 different classes such as Normal, MCI (Mild Cognitive Impairment) and Alzheimer’s Disease. The presented 3D CNN model can accurately classify Cognitive Normal patients and also MCI (Mild Cognitive Impairment) which may morph into Alzheimer’s Disease to prevent the further disintegration and severity of the patient’s condition. This paper also presents the comparative study of the performance of different Machine Learning models on the test diagnosis data of the subjects and suggests the most efficient approach to classify into 3 classes by data pre-processing and Hyperparameter optimization

Keywords: Alzheimer’s Disease, 3-D CNN, Hyper-parameter optimization, Machine Learning