Improving embedding efficiency of covering codes for applications in steganography

In image steganography, each pixel can carry a ternary message by choosing adding/subtracting one to/from the gray value. Although ternary covering functions can provide embedding efficiency higher than binary ones, it is necessary to convert the binary message into a ternary format. We propose a novel method that improves the embedding efficiency of binary covering functions by fully exploiting the information contained in the choice of addition or subtraction in the embedding. The improved scheme can perform, equally well with, or even outperform, ternary covering functions without ternary conversion of the message.