An Improved Fuzzy C-Means Clustering for Brain MR Images Segmentation

Abstract

In this paper, an improved fast FCM (HF-KFCM) algorithm was proposed based on histogram statistics of brain MR images. The algorithm firstly uses the multi-scale window traversal method to find the peak point of the histogram, then uses it as the initialization center of fuzzy clustering, and finally uses the fast clustering method based on statistical information to traverse, so as to reduce the computation amount of each iteration. Experimental results show that compared with the standard FCM algorithm and other improved algorithms, the proposed algorithm is significantly improved in clustering effectiveness and fuzzy segmentation.