On the detection of Alzheimer's disease using fuzzy logic based majority voter classifier

Alzheimer's disease (AD) is considered to be one of the most frequent neurogenerative dementia in the elderly population. In order to improve the quality of life span, early detection of AD has drawn significant attention to the researchers throughout the globe. To this aim, this paper makes a novel attempt to classify the brain MRI images into three classes viz. Alzheimer's disease (AD), mild cognitive impairment (MCI) and healthy control (HC) using the volumetric information of white matter (WM), grey matter (GM) and cerebro spinal fluid (CSF). This classification has been accomplished with the help of fuzzy logic based approach followed by a majority voter classifier. Our proposition is finally tested over several brain MRI images collected from ADNI dataset. Supremacy of our proposition has strongly been established by measuring its performance parameters such as accuracy, sensitivity and specificity and subsequently been compared with many of the state-of-the-art methods. Simulation results have shown an average improvement of approximately 6.5%, 6.9% and 4% over a number of existing works in terms of accuracy, sensitivity and specificity respectively.