A Machine Learning Based Approach for Early Detection of Alzheimer’s Disease by Extracting Texture and Shape Features of the Hippocampus Region from MRI Scans

Abstract: Alzheimer�s Disease (AD) is a neurological disease and the most common cause of dementia in the age group 65 years and above. The accurate and timely diagnosis of the AD is crucial in order to prevent the progression of this irreversible disease. This paper concentrates on a method to detect Alzheimer�s from MRI scans using machine learning approach. The proposed approach focuses on the hippocampal region of the brain. The texture features such as entropy, homogeneity, energy, contrast, correlation and variance are extracted from the hippocampus region using the Gray Level Co-Occurrence Matrix (GLCM). The area and shape feature are extracted using the Moment Invariants. The Error-back Propagation (EBP) in Artificial Neural Network (ANN) is used as the classifier for detection of various stages of AD. The proposed system gives an average accuracy of 86.8%.

Keywords: Alzheimer�s Disease (AD), Gray Level Co-Occurrence Matrix (GLCM), Artificial Neural Network (ANN).