A Fast Massively Parallel Fuzzy C-Means Algorithm for Brain MRI Segmentation

Abstract

In this paper, we propose a fast parallel algorithm for data classification, and its application for Magnetic Resonance Images (MRI) segmentation. The presented classification method is based on a parallel fine grained fuzzy C-means algorithm. It is implemented on a polymorphic SIMD machine to sort out the different components of a b rain image. The use of the massively parallel architecture in the classification domain and parti cularly for the fuzzy classification is introduced to improve the complexities of the corresponding al gorithms. The proposed algorithm is assigned to be implemented on a massively parallel machine, which is the Reconfigurable Mesh Computer (RMC). The brain image of size (m x n) to be proces sed must be stored on the RMC of the same size, one pixel per Processing Element (PE). Some i nteresting results are obtained in terms of accuracy and efficiency analysis of the proposed me thod, thanks to the reconfiguration ability of the used computational model.