Chapter 15 - Brain tissue segmentation using modified fuzzy C-means clustering

Abstract

This research proposes a new approach for segmenting and calculating tissue area in MRI images of the brain. Since the boundaries of the brain tissues are usually complex, we used K-means clustering, fuzzy C-means (FCM) and modified version of fuzzy C-means named as MFCM to segment the brain tissues into three different tissue segment namely white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF). The findings of the modified FCM (MFCM) segmentation are fed into the level set process for further refining. In this paper the work defines an efficient image segmentation approach using fuzzy C-means algorithm integrated with K-means clustering technique and it is followed by noise removal using median filter named a modified fuzzy C-mean (MFCM). We are here fixing the fuzzy parameter in such a manner which is actually giving the better result than the rest. Pixel counting is used to calculate the percentages of WM, GM, and CSF. This method can be used to calculate tissue volume in a variety of diseases and age groups.