A dynamic spatial fuzzy C-means clustering-based medical image segmentation

Abstract

This chapter presents a novel method for segmentation of normal and noisy MRI medical images. In the proposed algorithm, Genetic Algorithm-based FCM-S1 (GAFCM-S1) is used which takes into consideration the effect of the neighborhood pixels around a central pixel and exploits this property for noise reduction. While spatial FCM (SFCM) also considers this feature, this method is still preferable over the others as it is a relatively faster method. Moreover, GAFCM-S1 is self-starting and the centroids are completely independent of the user inputs. Turi’s Validity Index has been used as a measure of proper segmentation.