Segmentation of cerebrospinal fluid from 3D CT brain scans using modified Fuzzy C-Means based on super-voxels

Abstract

In this paper, the problem of segmentation of 3D Computed Tomography (CT) brain datasets is addressed using the fuzzy logic rules. In particular, a new method which combines Fuzzy C-Means clustering and the idea of super-voxels is introduced. Firstly, the method applies the extended Simple Linear Iterative Clustering (SLIC) method to divide image into super-voxels, which are next clustered by Modified Fuzzy C-Means algorithm. The method deals with 3D images and performs fully three dimensional image segmentation. Ten samples are supplied proving that our Modified Fuzzy C-Means (MFCM) together with super-voxels are apt to take into account a large diversity of special domains that appear and which are inappropriate solved adopting classical Fuzzy C-Means approach. The results of applying the introduced method to segmentation of the Cerebro-Spinal Fluid (CSF) from the brain ventricles are presented and discussed.