Effective weighted bias fuzzy C-means in segmentation of brain MRI

Abstract

Segmentation is a difficult task and challenging problem in the brain medical images for diagnosing cancer portion and other brain related diseases. Many researchers have introduced various segmentation techniques for brain medical images, however fuzzy clustering based fuzzy c-means image segmentation technique is more effective compared to other segmentation techniques. This paper introduces three new proposed algorithms namely Weighted Bias Field FCM [WBFCM], Modified bias field FCM [MBFCM] and New Approach to Bias field FCM [NBFCM] based on bias estimation and apply for segmentation of brain MRI. In general, the intensity in-homogeneities are endorsed to imperfections in the radio-frequency coils or to the problems connected with the image acquisition. The proposed methods are capable to deal the intensity in-homogeneities and more noised image effectively. We have compared our results with standard FCM and other reported methods. Further, to reduce the number of iterations, the proposed algorithms initialize the centroid using dist-max initialization algorithm before the execution of algorithm iteratively. The experimental results on brain MRI show that our methods is superior in providing better results compared to standard fuzzy c-means based algorithms.