AN APPROACH TO SEGMENT THE HIPPOCAMPUS FROM T2-WEIGHTED MRI OF HUMAN HEAD SCANS FOR THE DIAGNOSIS OF ALZHEIMER’S DISEASE USING FUZZY C-MEANS CLUSTERING

Abstract

The human brain plays a key role in memory-related functions such as encoding, storage, and retrieval of information. A defect in the brain results in memory impairment such as Alzheimer’s disease (AD). Atrophy in the volume of hippocampus (Hc) is the earlier symptom of AD. Therefore, to study the Hc, one needs to segment it from the magnetic resonance imaging (MRI) slice. In this paper, a semiautomatic method is proposed to segment the Hc from MRI of human head scans. The proposed method uses geometric mean filter for image smoothing. The fuzzy C-means clustering is applied to convert the filtered image into three distinct regions. From those regions, the image is classified into region of interest (ROI) pixels and non-ROI pixels. The proposed method is applied to five volumes of human brain MRI. The Jaccard (J) and Dice (D) indices are used to quantify the performance of the proposed method. The results show that the proposed method works better than the existing method. The average value of Jaccard and Dice is obtained as 0.9530 and 0.9744, respectively, for the five volumes.