期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2019
卷号：97
期号：24
页码：3661-3670
出版社：Journal of Theoretical and Applied
摘要：Intuitionistic fuzzy sets and rough sets are widely used for medical image segmentation, and recently combined together to deal with uncertainty and vagueness in medical images. In this paper, a rough set based intuitionistic fuzzy c-means clustering algorithm is proposed for segmentation of the magnetic resonance (MR) brain images. Firstly, we proposed Generalized intuitionistic rough fuzzy c-means [1] algorithm to overcome the dependency with the membership function, parameter tuning and histogram based initial selection of centroids to avoid local minima. In this paper, a modified generalized rough intuitionistic fuzzy c-means is proposed to avoid the histogram-based selection of initial centroids, instead from the obtained intuitionistic regions the centroids are calculated. Also, a minimized distance measure is proposed to improve the performance in all considered scenarios. Experimental results demonstrate the superiority of proposed algorithm.
