MRI BRAIN IMAGE SEGMENTATION USING HYBRID FUZZY CLUSTERING TECHNIQUE

Abstract

ABSTRACT Segmentation of human brain image from magnetic resonance imaging(MRI) into three main tissues such as gray matter(GM), white matter(WM) and cerebrospinal fluid(CSF) plays an important role in computer aided diagnosis and neuroscience research.It helps to detect various neurodegenerative and mental disorders. However, due to the presence of various artifacts like unknown noise, intensity inhomogenity and partial volume effect, the segmentation of MRI brain images is a challenging task. Possibilistic Fuzzy C-means algorithm(PFCM) is the hybridisation of fuzzy C-means(FCM) and possibilistic C-means(PCM) which is able to overcome the noise problem of FCM and the problem of coincident clusters in PCM.The major challenge that PFCM poses is to take into account the ambiguity in final final localisation of of feature vectors due to lack of qualitative information in the segmentation of ill defined MRI image with noise.This might result in improper assignment of memberships values ti the desired clusters.To overcome this, we have proposed a hybrid intuitionistic fuzzy C- means(HIFCM) clustering technique that hybrids PCM and FCM algorithms along with the use of intuitionistic fuzzy sets.Also. it includes a fuzzy factor to include local and spatial information and a novel possiblistic factor that maximises typicality. Real and simulated brain MRI images are segmented to show the superiority of our proposed HIFCM algorithm.The experimental results demonstrate that the proposed algorithm yields better results.