This paper present a steganalysis method for the data hiding scheme based on VQ-compression. This data hiding algorithm divides the codebook into groups which contain two codewords each. The Euclidean Distance of the group is used instead of a codeword in traditional VQ-compression. A PoV-like effect of Chi-square attack is observed and used as a feature of detection. In the proposed steganalysis, we detect whether an unknown image is a VQ-compressed image or not, and then the target detection of codewords grouping type data hiding methods is proposed. We apply proposed scheme to Yang et al.'s watermarking scheme. A large amount test image database UCID (Uncompressed Colour Image Database) is utilized as various conditions, such as cover imaged, traditional VQ-compressed images, and stego images. The experimental shows that the proposed steganalysis method is able to identify the stego images among others, and the accuracy rate reaches over 90 %. © 2014 IFIP International Federation for Information Processing.
CITATION STYLE
Chang, Y. T., Wu, M. H., & Wang, S. J. (2014). Steganalysis to data hiding of VQ watermarking upon grouping strategy. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 8407 LNCS, pp. 633–642). Springer Verlag. https://doi.org/10.1007/978-3-642-55032-4_65
Mendeley helps you to discover research relevant for your work.