期刊名称：International Journal of Computer Science and Information Technologies
电子版ISSN：0975-9646
出版年度：2016
卷号：7
期号：3
页码：1105-1110
出版社：TechScience Publications
摘要：Image segmentation is a challenging step inmedical image processing. It deals with separating thearea of interest from the images. Among the popularmethods of segmentation, clustering is an important onewhich clusters the image pixels based on the similarity ofpixels based on the centroid values. But clustering faces amajor problem of inhomogeneous clusters. In this paperwe propose a new segmentation algorithm called GainRatio based Fuzzy C-Means algorithm (GRBFCM) whichovercomes the problem of inhomogeneous clusters. ThisGRBFCM algorithm exploits the concepts of informationtheory in combination with fuzzy logic. The proposedalgorithm is applied to brain Magnetic Resonance Image(MRI) to segment tumor region. The segmentationaccuracy is evaluated using parameters like DiceCoefficient (DC ) and Jaccard Similarity (JS). We foundthat the proposed algorithm has better performance.
