Edinburgh Napier University

  In this article, we proposed a new scheme to encrypt highly autocorrelated image pixel data. In existing literature, single substitution was used to break autocorrelation in images. To get better results, instead of single substitution box, some researchers are minimizing the autocorrelation via utilizing multiple substitution boxes. However, we found that multiple substitution boxes cannot solve the problem of autocorrelation independently. Therefore, we added chaotic confusion and diffusion to the existing substitution scheme in order to remove the correlation completely. Initially, we break the autocorrelation of data by permuting the pixels via TD-ERCS chaotic map. Bitwise XOR is carried out for diffusion using logistic map followed by the existing substitution scheme. Experimental analysis shows that the proposed scheme has greater resistance to brute force attack and statistical attack. Further analyses such as statistical analysis, peak signal-to-noise ratio, contrast analysis, key space analysis, entropy analysis, correlation analysis, key sensitivity analysis, and encryption quality analysis proved the high robustness of the proposed scheme.