A comparative study of cellular automata-based digital image scrambling techniques

Cellular automata (CA) are an important class of dynamic systems, discrete in both time and space units. A cellular automaton evolves by the local interaction of its discrete space units or cells at discrete time steps. This local interaction is governed by simple rules that compute the next state of each cell. Many of these rules evolve CA to generate chaotic or complex patterns and, as such, these CA rules find application in a wide variety of areas including digital image scrambling (DIS). The dynamic behavior of any given CA is largely influenced by the non-quiescent state ratios present in the initial CA configuration. In this paper, we first implement and analyze different CA-based DIS techniques using same parameters, wherever possible, and same dataset of test images for a justified comparison of their performance in terms of gray difference degree (GDD). Next, the effect of different non-quiescent state ratios in the initial CA configuration, and varying image sizes on GDD using these CA-based DIS techniques is analyzed. Robustness of all the DIS techniques is evaluated using correlation coefficient analysis and number of pixels change rate.

Automated summary

This paper implements and analyzes different cellular automaton-based image scrambling techniques to compare their performance in terms of gray difference degree (GDD). The study also evaluates the effects of different non-quiescent state ratios in the initial CA configuration and varying image sizes on GDD using these techniques. The robustness of all the techniques is assessed through correlation coefficient analysis and number of pixels change rate.