A new binary similiraty metric for two LSB steganalysis
Lakhdar Laimeche, Hayet Farida Merouani, Mazouzi Smain · 2016
In this paper, a novel steganalytic method is proposed for 2LSB (Two Least Significant Bits) steganography. This method is original from two perspectives: it is based on the application of Zipf's law as an analysis tool for images, and it proposes a new binary similarity metric (BSMC) between the seventh and eighth bit planes. Because the hidden message leaves statistical artifacts on the cover image, the embedding process often decreases the correlation existing in bit planes of an image to some extent. Therefore, Zipf's law is used to compute the binary similarity metric between curves of Zipf of the cover and stego images, which is based on the distribution of patterns of the 7thand 8thbit planes. This method can not only detect the presence of secret messages embedded using 2LSB embedding, but it can also be easily generalized to multi least significant bit embedding (MLSB).