Evolutionary-based image encryption with dna coding and chaotic systems

Abstract

In order to obtain encrypted image with optimal correlation coefficient, we propose a new evolutionary-based image encryption scheme in this paper, based on DNA coding, chaotic systems, and a genetic algorithm (GA). Firstly, the DNA encoding operation is performed over the original plain image. Then logistic mapping and hyper-chaotic systems are used to create the number of encrypted images over the DNA matrices. We fix these encrypted images as the initial population for the GA, and the GA is applied to them to determine the best one based on the fitness function. The correlation coefficient is used to define the fitness function in this paper. Finally, after the evolution, the encrypted image with the lowest correlation coefficient will be obtained as the final result. The novelty of this research is in using the logistic map over the DNA matrix of the original image directly, which will make the initial group of encrypted images more secure and with lower initial fitness for the GA. Therefore, the scheme will achieve high fitness in fewer iterations and retain the efficiency of the encryption. Experimental results confirm that the proposed scheme not only has a good encryption effect, but also has the ability of resisting statistical analysis attacks, brute force attacks, and differential attacks.