Steganalysis of Adaptive Steganography Under the Known Cover Image

Jialin Li, Xiangyang Luo, Yi Zhang, Pei Zhang, Chunfang Yang, Fenlin Liu

Abstract: Digital image steganography is one of the important ways to achieve covert communication to protect users’ data privacy. Steganalysis technology is the key technology to check the security of steganography, and the extraction of embedded messages is an important challenge for image steganalysis technology. Existing methods are for plain text embedding or only for certain special scenarios, they are not applicable to the extraction under the known cover image. To this end, this paper proposes a method of extracting embedded messages under the condition of the known cover images. First, the STCs encoding process is analyzed using the syndrome trellis. Second, a path in the syndrome trellis can be obtained by using the stego sequence and a certain parity-check matrix. Meanwhile, the embedding process can also be partially simulated using the cover sequence and the parity-check matrix. By comparing whether the paths are consistent, the coding parameters can be quickly filtered to find the correct submatrices and extract the correct embedded messages. This algorithm avoids the second embedding of all possible secret messages, which significantly improves the efficiency of coding parameter recognition. The experimental results show that the proposed method can identify STCs parameters of stego images using HUGO steganography in a short time, so as to realize the extraction of embedded messages.