A Novel Image Encryption Algorithm Based on Double Permutation and Random Diffusion

Abstract To improve the image transmission security, an image encryption algorithm based on double permutation with random diffusion is proposed in this paper. This algorithm consists of three stages: adaptive key generation, double scrambling and random diffusion. Firstly, in order to make the key highly correlated with the plaintext image, we use the SHA-256 value of the plaintext image to generate the key. Then, the double permutation is mainly composed of pixel-level permutation and bit-level permutation. The pixel-level permutation uses chaotic sequences generated by the chaotic system to generate the permutation sequences and uses these two sequences to rearrange the row and column coordinates of the image. Secondly, we implement bit-level permutation using the Fisher-Yates Shuffle algorithm as a group of every four neighboring pixels. The double permutation breaks the high correlation between image pixels. Thirdly, in the diffusion stage, we propose a random diffusion algorithm to address the problem that the diffusion order and direction of current diffusion methods are fixed and not random. The permuted image is divided into four blocks, and the order and direction of the diffusion operation for each pixel in each block is determined according to the chaotic sequence from the chaotic system. Finally, experimental results and security analysis show that the algorithm has good performance and sufficient security. It can resist a variety of attacks, such as statistical attacks, brute force attacks, and plaintext/selective plaintext attacks.