Abstract:

Nowadays, some image encryption methods adopt the scrambling algorithm and low-dimensional chaotic system that have the inherent features of small key space and low complexity, which makes the algorithm vulnerable to chosen plaintext attack. In this paper, a color image encryption method using the genetic simulated annealing algorithm and chaotic systems is proposed to achieve a better security performance. First, the plain image is processed by the selection and crossover operations. Then, the optimal pseudo-random sequences are generated to scramble the image based on the simulated annealing algorithm. These three sets of operations make the histogram of the scrambled image uniform, which can resist the statistical attack. Finally, in order to enhance the correlation of each component of the image, the interactions among multiple components are used to mutate the scrambled image, and the mutation operation is judged by the fitness of the plain image and scrambled image. Compared with the classical encryption architecture, the proposed method not only increases the complexity of the cryptosystem, but also enhances the sensitivity of the encryption method to the plain image. Experimental results and performance analysis show that the proposed method has a large key space, high security and high sensitivity to the plain image, which can resist common cryptanalysis attacks effectively.

Key words: image encryption, chaotic systems, genetic simulated annealing algorithm, color image