A fast ARX model-based image encryption scheme

This paper proposes a novel ARX model-based image encryption scheme that uses addition, rotation, and XOR as its confusion and diffusion mechanism instead of S-Box and permutation as in SP networks. The confusion property of the proposed scheme is satisfied by rotation and XOR with chaotic sequences generated from two logistic maps. Unlike classical image encryption schemes that adopt S-Box or permutation of the entire plain image, the diffusion property is satisfied using addition operations. The proposed scheme exhibits good performance on correlation coefficients (horizontal, vertical and diagonal), Shannon's entropy and NPCR (Number of Pixels Change Rate). Furthermore, simulation results indicate that its time complexity is 9.2 times more efficient than the fastest algorithm(Yang's algorithm).