PSO optimized Pulse Coupled Neural Network for Segmenting MR Brain Image

Abstract

In this proposed method, Magnetic Resonance (MR) Brain image segmentation technique based on Pulse Coupled Neural Network (PCNN) clustering combined with Particle Swarm optimization (PSO) approach has been presented. Since, PCNN is robust to noise, the input image is added with 0.05 Level of impulsive noise and the segmented output was analysed based on the fractions, selectivity and sensitivity. Accuracy of the proposed technique was found to be 93%. Moreover, in this proposed method, instead of selecting the parameters of PCNN in a random manner, they are optimized using PSO technique.