Chaotic map-embedded Blowfish algorithm for security enhancement of colour image encryption

Abstract

The quest for achieving an accurate and efficient algorithm in enhancing the security of colour image encryption is never-ending. We enhance the security level of conventional Blowfish algorithm (BA) for colour image encryption by modifying it with new F-function. The dynamic S-box and XOR operator are generated from the F-function via four-dimensional hyper-chaotic map, and the number of iterations is dramatically reduced from sixteen to four to simplify the processing complexity. The block secret keys of varying space size are randomly created. The computation is performed using Matlab R2008a codes on eight images of size 512×512 pixels obtained from standard USC-SIPI Image Database. The security of the proposed method exhibits notable enhancement with increasing block size. Performance analyses reveal a correlation coefficient less than -0.0016 and entropy greater than 7.9993. Furthermore, the distribution of RGB channel for image Lena displays sizeable security improvements. Correlation between image pixels is significantly decreased with higher entropy. In comparison with previous works, the proposed algorithm is found to exhibit superior performance for colour image encryption than the conventional BA.