A novel image encryption scheme based on ccnn

Abstract

Abstract With the advancement of computational capacity, the key space will become one of the crucial factors influencing the security of digital cryptographic systems. Despite chaotic-based digital cryptographic systems possessing large key spaces, the post-Moore’s era rapid growth in computational capacity continues to pose challenges to the security of chaotic-based cryptographic systems. To address this issue, a novel image encryption scheme based on non-autonomous chaotic system is presented in this paper. In particular, a brain inspired neuron called continuous-coupled neural network (CCNN) is utilized to design image encryption scheme. To achieve the efficient image encryption scheme, firstly, the CCNN model is simplified to uncoupled-linking neuron model. The dynamic behavior under various driving signals is studied. The analysis showed that uncoupled-linking CCNN neuron exhibit various dynamic behavior under sine waves, triangular waves, sawtooth, superimposed sine waves, etc. Secondly, the decorrelation operation method is utilized to enhance the pseudo-randomness of the sequence. On this basis, thirdly, the image encryption scheme is proposed. It uses bit-level pixel scrambling, row scrambling, column scrambling and diffusion to modify the pixel value and the pixel position of the image. Security analysis shows that the proposed scheme is able to resist differential attack, statistics attack, known-plaintext attack and brute force attack. Moreover, the key space of the proposed scheme can be extended by the combination of drive signals. This unique feature makes the key space of the proposed scheme to be infinite, leading this kind of chaos-based cryptographic system to be a competitive candidate in post-Moore’s era.