A Hybrid Approach for Segmenting Grey and White Matter from Brain Magnetic Resonance Imaging (MRI)

Abstract

Magnetic Resonance Imaging (MRI) is a common medical imaging diagnostic tool for the identification of disease(s) during clinical investigation. Brain MRI is used for diagnosis of brain-related diseases such as brain tumours, Alzheimer’s disease, etc. This has proven to be advantageous over other diagnostic techniques and also adds to the versatility and diagnostic utility for surgical treatment planning and clinical interventions. Brain tissues have grey matter and white matter whose intensity is almost similar, hence making the diagnosis of the brain-related disease difficult. Segmentation of grey matter and white matter is crucial to detect the various brain-related disease such as Alzheimer’s, Migraine, Huntington, Multiple sclerosis and Dyslexia, which show significant volumetric changes in grey matter and white matter. Prior to the segmentation of brain regions, skull stripping is a necessity for accurate diagnosis of various brain-related diseases. In this paper, histogram- based skull stripping technique is applied to separate the skull and then a novel hybridised technique is proposed using Fuzzy Edge Detection and Region-Growing to Segment the Grey and White Matter from Brain MRI. The result of the proposed technique is compared with different existing techniques such as Region growing; Histogram-based method, fuzzy C- Means, K Means, etc. It is found that the proposed method produces convincing results.