A New Alzheimer's Disease Classification Technique from Brain MRI images

Alzheimer's Disease (AD) is one of the most common forms of dementia that leads to memory loss mostly in elderly people. Treatment at the early stage of the disease will lessen the progression rate of the disease and its pathogenesis. Computer Aided Diagnosis (CAD) is gaining popularity in recent years for the detection of AD. In this study, we report a novel strategy for AD classification using brain Magnetic Resonance (MR) Images. Ensemble Empirical Mode Decomposition (EEMD) was used to decompose the image. From each decomposed image, the statistical features and fractal based features were extracted. Feature reduction has been performed using Principal Component Analysis (PCA). The proposed method was examined on publicly available data sets (OASIS). Three different classifiers namely Logistic regression, NaiveBayes, and support vector machine was used in this work. The classification accuracy of each classifier was reported and compared with the existing methods. The average accuracy achieved was 92.34%. The developed approach seems to be promising to develop reliable and accurate devices for AD detection, and classification.