Fuzzy VQ based image compression by bacterial foraging optimization algorithm with varying population

Abstract

In this paper, a new variant of bacterial foraging optimization (BFO) algorithm, called bacterial foraging optimization algorithm with varying population (named BFVPA) is proposed for Fuzzy Vector Quantization based image compression. The work shows how BFVPA can be effectively utilized for reduction in average distortion measure between training and reconstructed image and how it can improve upon the performance of BFOA utilized for solving similar problems. In contrast to BFOA, where a fixed population of bacteria is utilized, the basic essence of BFVPA is that the population size undergoes variation through the phases of chemotaxis, metabolism, elimination and quorum sensing, in each iteration. The proposed algorithm has been employed on several benchmark gray scale images and the compression performances are computed in terms of a popular performance index, called PSNR. The performances show that BFVPA is able to provide an overall, superior performance compared to that of BFOA.