Robust algorithm for early detection of Alzheimer's disease using multiple feature extractions

• Main Authors: Mathew, Jesia, Mekkayil, Lasitha, Ramasangu, Hariharan, Karthikeyan, B. Ramaswamy, Manjunath, Aditya G.
• Format: Conference Proceeding
• Language: English
• Subjects: Alzheimer's disease; Alzheimer's disease (AD); Discrete Wavelet transform(DWT); Discrete wavelet transforms; Feature extraction; Fisher Discriminant Ratio (FDR); Magnetic resonance imaging; Mild Cognitive Impairment(MCI); Principal component analysis; Principal Component Analysis(PCA); Support Vector Machine(SVM); Support vector machines
• ISSN: 2325-9418
• DOI: 10.1109/INDICON.2016.7839026

An algorithm is proposed for early detection of Alzheimer's disease and is focused on detecting the condition that would lead to Alzheimer's disease in future. Alzheimer's disease is a prevalent case now and it mostly affects the elderly people. The disease condition makes a person lose his memory and have trouble in doing his day-to-day activities, and progressively the condition leads to death. No treatment is available to completely cure the disease but it would be beneficial if the disease is detected earlier as the necessary aid can be provided. The proposed algorithm uses Alzheimer's Disease (AD), Cognitively Normal (CN), Mild Cognitive Impairment non-convertible (MCInc) and Mild Cognitive Impairment convertible (MCIc) data from the ADNI (Alzheimer's Disease Neuroimaging Initiative) and it involves the process of pre-processing, feature extraction and classification. MCIc leads to Alzheimer's disease at a later stage. A combination of the Discrete Wavelet Transform (DWT) and the Principal Component Analysis (PCA) is used for the feature extraction and further subjected to classification using the Support Vector Machine (SVM). The multiple feature extractions with both the DWT and the PCA together give a better accuracy during classification compared to other algorithms.