Diagnosis of Alzheimer disease using fast independent component analysis and Otsu multi-level thresholding

Abstract

Detection of Alzheimer disease using Magnetic Resonance Imaging (MRI) is the most challenging aspect in the field of medical image processing and analysis. In this paper, the proposed methodology has three major steps: image acquisition, image pre-processing and segmentation. Initially, the brain images were acquired from the dataset: Open Access Series of Imaging Studies (OASIS). After image acquisition, image pre-processing was carried out using median filter, it utilized for cutting down the noise and to improve the quality of acquired brain images. Then, segmentation was carried-out using Fast-Independent Component Analysis (Fast-ICA) along with Otsu multilevel thresholding. It was a flexible high level machine learning technique to localize the object in complex template. In experimental analysis, the proposed approach distinguishes the brain MRI tissues: White Matter (WM), Cerebro- Spinal Fluid (CSF), and Grey Matter (GM) by means of Tanimoto index, similarity index, precision, and recall. The proposed methodology improved the Alzheimer tissue detection up to 15-30% compared to the existing methods: Band Expansion Process (BEP), ICA and BEP-ICA in terms of precision and recall.