Modeling hyperchaotic datasets for neural networks

Abstract

This work is aimed at the studies related to neurocryptography. The paper represents the studies of hyperchaotic mappings and their construction based on the attractors and the research of image noise characteristics using the attractors and their performance. The conducted experiments have demonstrated that Liapunov hyperchaos generator possesses the best performance ratio and noise characteristics. In prospect we are going to conduct the experiments with a compiled data set and neural networks focused on the work with chaotic models and cryptographic algorithms.