A Steganographic Embedding Undetectable by JPEG Compatibility Steganalysis
FLORIDA UNIV GAINESVILLE COMPUTER INFORMATION SCIENCE AND ENGINEERING
Pagination or Media Count:
Steganography and steganalysis of digital images is a cat-and-mouse game. In recent work, Fridrich, Goljan and Du introduced a method that is surprisingly accurate at determining if bitmap images that originated as JPEG files have been altered and even specifying where and how they were altered, even if only a single bit has been changed. However, steganographic embeddings that encode embedded data in the JPEG coefficients are not detectable by their JPEG compatibility steganalysis. This paper describes a steganographic method that encodes the embedded data in the spatial domain, yet cannot be detected by their steganalysis mechanism. Furthermore, we claim that our method can also be used as a steganographic method on files stored in JPEG format. The method described herein uses a novel, topological approach to embedding. The paper also outlines some extensions to the proposed embedding method.