Classification of Brain MRI for Detection of Alzheimer’s Disease

Abstract

In this study, we propose a method for the classification of T1-weighted Magnetic Resonance Images (MRI) of the Brain in case of Alzheimer’s disease based on extraction of different features. The dataset used comprises of right handed females of age group 18-96 years and is helpful to determine the onset and early detection of the disease. The processed MR images are used to obtain different features for training of classifiers. The whole Grey matter is considered as the Region of Interest (ROI) because it contains the Amygdala and Hippocampus, the regions most affected due to Alzheimer’s. The features used for this study are some of the crucial second order features derived from Grey Level Co-occurrence Matrix (GLCM) such as Entropy, Energy, Homogeneity and Correlation and also the ratio of the Grey Matter Volume and the White Matter Volume to the Volume of the Cerebrospinal Fluid. An accuracy of 84% has been achieved with a sensitivity of 100%. This proves to be a better method than Voxel Based Morphometry extraction method which is cumbersome but has been proved to be more accurate but not as sensitive.