A method for designing S-box based on chaotic neural network

S-box (Substitution box) is one of the most important components in the block cipher. In this paper, the main indices to evaluate S-box are firstly introduced. Then an algorithm for obtaining cryptographical strong S-boxes based on chaotic neural network is proposed. The chaotic neural network used to design S-boxes is composed of two layers, which have sixteen and eight neurons, respectively. Owing to making full use of the nonlinearity of the chaotic network, some S-boxes with good cryptographical performance can be produced. Therefore, the proposed method is a promising choice for designing S-boxes and suitable for designing block cryptosystem with dynamical S-boxes.