CNN-based color image encryption algorithm using DNA sequence operations

Abstract

The encryption algorithm of color images based on chaos theory has attracted lots of attentions in recent years. Nevertheless, due to the defects of the low dimensional chaotic system in single structure and small key space size, the security of cryptosystem is not sufficient enough. In this paper, we proposed a novel encryption algorithm for color images based on Deoxyribonucleic acid (DNA) sequence operations and cellular neural network (CNN) to dispose of these defects. The proposed cryptosystem of this paper takes on the features of large key space and complex structure. Firstly, the plain color image is split into three matrices (R, G, B) which are transformed into DNA matrices by the DNA encoding rules, respectively. Secondly, the elements' positions of the three DNA sequence matrices are scrambled via the chaotic sequences generated by CNN. Thirdly, the three DNA matrices are summed according to the certain rules and complemented by the complementary rules, and then the cipher-image is obtained by the DNA decoding rules via the DNA matrices. Simulation results and security analysis show that the encryption effect of this paper is not only better than traditional encryption algorithms but also has excellent ability to hold back familiar attacks.