Digital Image Encryption Algorithm Based on Multi-Dimensional Chaotic System and Pixels Location

Abstract

Any image encryption system divided mainly into two Methods: pixel replacement methods and pixel scrambling methods. In the pixel replacement method, each pixel in the image needs to change its value. Where, in the scrambling method the pixel needs to change its position. In this paper, we propose a new image encryption algorithm based on multi-dimensional chaotic function called a Rossler attractor; to enhance the encryption system, increase the complexity of the encryption keys and decrease the computational complexity of the cipher image. To do that, the Rossler attractor with the three dimensional planes as the encryption keys for the first level and the second level encryption are used, where, to decrease the execution time and computational complexity we used only one non-linear term function. Furthermore, our algorithm consists of two scrambling methods and two replacement methods. In which, each value in the image will be replaced by using a XORing operation with its location and change its location by using two shuffling approaches. Moreover, 3-planes chaotic function was used to scramble the pixels position as follows: in the first method we used X, Y planes of the Rossler attractor and in the second method we used X,Y and Z planes, so, the uncertainty of the adjacent pixels will be increased. However, by analyzing our algorithm, we show that the key space of our algorithm is equal to 10 45 , where, the entropy tests shown that the average entropy for the tested images is not less than 7.9971.Furthermore, after analyzing the histogram and correlation between the adjacent pixels, we show that our algorithm is strong against different types of attacks and it's sensitive to the initial conditions.