A New Scrambling Method Based on Arnold and Fermat Number Transformation

In the preprocessing of digital image information hiding, the common scrambling method Arnold based on locational space canpsilat change the statistical character of an image. Because of using fixed form matrix, the image is detected and attacked easily. Combining with the Fermat number transformation matrix changeable, the fast algorithm like FFT and solving the inverse matrix simply, this paper presents a new secondary scrambling method based on color space and locational space. Experiments show this method based on Arnold can reach ideally scrambling effect after only one time Fermat number transformation, and it is robust to cropping attack, so it is very useful in digital image information hiding.