Novel hyperchaotic image encryption method using machine learning-RBF

In this paper, we put forward a novel hyperchaotic image encryption method using machine learning-RBF. First, a new 4D continuous hyperchaotic system is designed to address the degradation issue of low-dimensional continuous chaotic systems, which has a simpler structure, wider chaotic range, better distribution, and higher complexity compared with other chaotic systems based on Hopfield-type neural networks. Additionally, it has good randomness and can be implemented using hardware-based digital signal processing. Then, based on this system, we explore a new image encryption method using machine learning-radial basis function (RBF) neural network and true random numbers. Results show that compared with some other algorithms, our method is more secure and withstand common attacks.