Improved MR Brain Image Segmentation Using Adaptive Gabor Filtering Scheme with Fuzzy C-Means Algorithm

Abstract—Image segmentation is the foremost process in medical image processing. It aids the diagnostic and clinical analysis of MRI (Magnetic Resonance Imaging) images that were acquired through the most complex procedures of medical diagnostics. The earliest soft computing techniques in segmenting images were carried out through Fuzzy C-Means (FCM) and similar extensions of various clustering algorithms. In this paper, we introduced an innovative method that uses Gabor energy filter with adaptive features to pre-extract the information of various regions of a brain image, obtained either from a MRI or CT scanner. The noise-reduced image with blurred features was then made to undergo modifications by applying unsupervised learning methods such as FCM technique, whose output has efficient exclusion of certain strength of noise elements resulting in better classified pixels