Title: Differential evolution versus genetic algorithm in optimising the quantisation table for JPEG baseline algorithm
Authors: Balasubramanian Vinoth Kumar; Manavalan Karpagam
Addresses: Department of Computer Science and Engineering, PSG College of Technology, Coimbatore – 641004, Tamil Nadu, India ' Department of Computer Science and Engineering, PSG College of Technology, Coimbatore – 641004, Tamil Nadu, India
Abstract: Quantisation table used in joint photographic experts group (JPEG) standard plays a major role in the trade-off between the compression and quality of the image. Hence, finding a quantisation table is inferred as an optimisation problem. This paper proposes a classical differential evolution (CDE) algorithm for finding the optimal luminance quantisation table for a target bits/pixel (bpp). Also, it aims to carry out an extensive performance comparison of the CDE algorithm with the classical genetic algorithm (CGA). It is observed that the CDE-based quantisation table outperforms the default JPEG and CGA-based quantisation tables in terms of mean squared error (MSE) and peak signal to noise ratio (PSNR). Images with different complexity levels have been experimented with both CDE and CGA algorithms. The validation results show that CDE guarantees a feasible solution in faster convergence rate than CGA. The results are confirmed using the statistical hypothesis test (t-test).
Keywords: image compression; JPEG baseline algorithm; quantisation table; optimisation; metaheuristics; genetic algorithms; differential evolution; statistical hypothesis test; t-test; image quality; mean squared error; MSE; peak SNR; signal to noise ratio; PSNR.
International Journal of Advanced Intelligence Paradigms, 2015 Vol.7 No.2, pp.111 - 135
Received: 30 Aug 2014
Accepted: 19 Jan 2015
Published online: 24 Jul 2015 *