Y. Wang, Y. Cao, X. F. Zhao
This paper presents a novel steganalytic method for detection of quantized DCT-based video steganography. First, the modification on the partially decoded quantized coefficients is modeled. Then the influence of the embedding operation on spatial domain is illustrated, which takes the form of the centralization of errors within each corresponding spatial pixel block. Finally, based on this fact, a 36-dimension feature set is extracted and used for classification. Experiments are carried out on videos corrupted by various quantized DCT-based steganography methods and encoded by various motion estimation methods. Performance results demonstrate the effectiveness of our scheme.
Cite this paper as:
 Y. Wang, Y. Cao, X. F. Zhao, Video Steganalysis Based on Centralized Error Detection in Spatial Domain. The 12th International Conference on Information Security and Cryptology (InsCrypt 2016), Nov. 4-6, 2016 Beijing, China, LNCS 10143: 472-483, Springer. [PAPER]