S-box design based on optimize LFT parameter selection: a practical approach in recommendation system domain

The strength of cryptography encryption is strongly depended upon the substitution box (S-box), hence highly non-linear S-box is appreciated in cryptography. In this work, we proposed a scheme to select the optimized values for control parameters (a, b, c, d) of linear fractional transform (LFT) to achieve an S-box with non-linearity of 112. For dynamic parameter values, we used location of chaotic sequences. The proposed s-box tested on different criteria such as non-linearity method (NL), probability methods (linear approximation (LP), differential approximation (DP), strictly avalanche criteria (SAC) and bit independence criteria (BIC). Our result shows that proposed s-box achieves better or equal to cryptographic strength as compared with state-of-the-art techniques. Moreover, we tested the proposed s-box based encryption method in recommendation system scenario with the idea of text-to-image conversion. The statistical analysis of this image encryption shows the bright prospects of the proposed method.