Design of double-coupled HR-FN neural network with memristors and its application in image encryption

In this paper, a double-coupled neural network based on hyperbolic memristors between HR neuron and FN neuron is proposed. Complex dynamic behaviors of the neural network are investigated by stability analysis and dynamic analysis. The maximum Lyapunov exponent cube of three-parameter is proposed, which is used in the dynamics analysis. The Lyapunov exponent cube allows a more intuitive observation of the dynamic characteristics of neural network. The accuracy of numerical analysis is validated through the circuit implementation of neural network. Furthermore, a medical image encryption algorithm based on double-coupled neural network is proposed. The image encryption algorithm combines Latin square permutation and Bit cross-coded diffusion, which effectively improves the randomness of the pixel replacement process. A series of random tests are used to detect the algorithm. Simulation results and numerical analysis show that the algorithm has high security and anti-cracking ability.