CHAOTIC IMAGE ENCRYPTION WITH HOPFIELD NEURAL NETWORK
Yuwen Sha, Jun Mou, Jue Wang, Santo Banerjee, Bo Sun
Abstract
With modern cryptography evolves, some sensitive information needs to be protected with secure and efficient algorithms. In this context, we found that Hopfield neural network (HNN) has stronger memory and can generate luxuriant kinetic behavior, especially with the introduction of fractional-order operators. Therefore, we propose a chaotic image encryption based on the fractional-order HNN (FO-HNN), where FO-HNN appears as a key generator. To de-correlate the correlation between pixels, a spatial permutation strategy is designed first, and then a new diffusion technique based on a Three-input logic valve is adopted to guide the diffusion process. Simulation results and security analysis show that the HNN-based image cryptosystem has superior security performance.