Matrix Embedding With Pseudorandom Coefficient Selection and Error Correction for Robust and Secure Steganography

摘要: In matrix embedding (ME)-based steganography, the host coefficients are minimally perturbed such that transmitted bits fall in a coset of linear code, with syndrome conveying hidden bits. The corresponding distortion and vulnerability to steganalysis significantly less than conventional quantization index modulation (QIM)-based hiding. However, ME is robust attacks, single bit error leading multiple decoding errors for this paper, we employ ME-RA scheme, combination ME-based hiding powerful repeat accumulate (RA) codes correction, address problem. A key contribution paper compute log likelihood ratios RA decoding, taking into account many-to-one mapping between an encoded bit, ME. To reduce detectability, hide randomized blocks, as recently proposed Yet Another Steganographic Scheme (YASS), replacing QIM-based YASS by scheme. We also show performance can be improved employing punctured codes. Through experiments based on couple thousand images, same embedded data rate moderate attack level, method results lower detection obtained YASS.