Image Encryption Algorithms Based on Non-uniform Second-Order Reversible Cellular Automata with Balanced Rules

Abstract

In this paper, we propose an image encryption algorithm based on non-uniform second-order reversible cellular automata (CA) which consist of two-state cells and use the Moore neighborhood. Different from other existing algorithms based CA, our algorithm uses the balanced CA rules that have a higher randomness than unbalanced rules. More specifically, before encryption we first mix the pixels of the plain image and then group these pixels with 128 bits for each block. Then each block is encrypted based on non-uniform second-order reversible CA by an iterative method. Since different key are constructed and applied for each block, the properties of confusion and diffusion of the encryption algorithm are well satisfied, and our algorithm can avoid the problem in the Electronic Codebook Book (ECB) encryption method that two plaintext blocks that have identical content are encrypted into identical ciphertext blocks. Simulation results show that our algorithm performs well in several security measurements such as histogram, information entropy, correlation of adjacent pixels, and sensitivity.