Steganalysis to adaptive pixel pair matching using two-group subtraction pixel adjacency model of covers

Steganalysis has played a positive role on protection and improvement of information security. The adaptive pixel pair matching (APPM) embedded stego signal which is independent in any notational system by providing more compact neighborhood and with lower distortion. No method is known to be applicable to estimation of the APPM. In this paper, a novel steganalysis scheme is presented to effectively detect the APPM steganography. Based on the subtractive pixel adjacency model of covers (SPAM), two-group modeling differences between adjacent pixels along horizontal, vertical, and diagonal directions are used to enhance changes caused by APPM steganography using a second-order Markov chain. Subsets of sample transition probability matrices are then used as features classified by support vector machines. The experiment results demonstrate that the proposed method is efficient to detect the APPM steganography compared to SPAM.