An efficient JPEG steganographic scheme using uniform embedding

Abstract

Steganography is the science and art of covert communication. Its objective is to hide the most secret messages into a cover object with the least possible statistical detectability. In practice, this is generally realized by a framework of minimal distortion embedding. This paper presents an efficient JPEG steganographic scheme based on syndrome trellis coding (STC) and a uniform embedding strategy, which, instead of random modification, tries to modify nonzero quantized DCT coefficients of different magnitude with equal probability, leading to possible minimal artifacts for statistics of DCT coefficients as a whole. The distortion metric corresponding to the uniform embedding is based on the magnitude of the DCT coefficients and both their intra- and inter-block neighborhood coefficients and known as uniform embedding distortion metric (UED). With the proposed scheme, the STC provides multiple codewords for a given message, while the UED determines the best one with minimal distortion. In this way, the average statistics change in each bin is significantly reduced, which corresponds to less detectability of steganalysis. Compared with prior arts, experimental results demonstrate the superior performance of the proposed scheme in terms of secure embedding capacity against steganalysis.