Cryptographic mixing transformations for image applications

Several image encryption methods have been recently considered. This paper presents a new image enciphering method starting from the mixing transformations suggested by C.E. Shannon for natural language. There are two ways in which images are involved here: first, as a means of visual perception of the mixing transformation features (the very sophisticated diffusion and confusion); secondly, as the application field itself. Both the theoretical and experimental aspects pointed out here show the high quality of this method of image enciphering. The method illustrates were made either for simulated images with different probability laws and autocorrelation functions, or for natural images.