FSR과 CA 기반의 영상 암호화

Abstract

In this paper, we propose a new method for image encryption based on the principles of FSR(Feedback Shift Register) and CA(Cellular Automata), in order to improve the security for image data that become increasingly important in information protection area. FSR generates a sequence with the maximum cycle, and is often used in bit stream encryption. CA deals with the time and space in a discrete manner, where the status each cell can have is simultaneously renewed by local interactions. MLCA(Maximum Length Cellular Automata) can generate the PN(pseudo noise) sequence with the maximum cycle at a minimal cost, and will add noise to the target domain. On the other hand, CAT(Cellular Automata Transform) is a method that transforms a physical domain into a CA domain, using basis functions and conversion formulas generated by Gateway Values. In this paper, we propose the image encryption method by applying the characteristics of FSR and Two-Dimensional CAT in a stepwise manner. FSR will generate a sequence with the maximum cycle that will be used as the basis image. This image will be converted through an XOR operation with the original image. Then, a Two-Dimensional CAT basis function will be generated and multiplied to the converted image to give a high-level encrypted image. Being a bit stream method, this method provides a quick encryption and high security. Also we propose the image encryption method that applies MLCA, which has a boundary condition of NBCA and IBCA, and Two-Dimensional CAT in a stepwise manner. The MLCA method is processed in the following order. First, a transition matrix is created using the Wolfram Rule. Then, the transition matrix is in operation to the original image that is intended to be encrypted, which transfers the pixel values of the original image. Furthermore, a Two-Dimensional CAT basis functions are generated and applied to the converted image in order to encrypt it. Finally, in order to verify the proposed method, we have performed stability analysis with respect to CA key space, pixel sensitivity and the histogram for general images, high-quality medical images, Web text, map images, and etc. As a result, it was confirmed that the methods could achieve a high encryption level. This method showed verification of no loss encryption method, which has restored the original images perfectly.