We propose a novel image encryption algorithm based on the stochastic Vasicek process and a modified chaotic logistic map. The proposed algorithm relies on the Fridrich confusion-diffusion structure. In the initialization step, we generate process parameters and an initialization state from a secret key of arbitrary length. Then the Vasicek process performs several iterations, where the number of iterations is generated at each step by the secret key. The floating point numbers generated by the process are discretized to generate the permutation for pixel confusion, and then generate pseudo-random sequences to diffuse the pixel values. The confusion-diffusion process can be applied for several rounds to improve the encryption performance. Experimental results and security analysis show that our model has good performance. The algorithm has a large key space, passes statistical analysis and has a strong resistance to differential attacks, which confirms that our algorithm is safe and efficient.
image encryption, Vasicek process, logistic map, diffusion, confusion
65H05, 65F10