Robust adaptive steganography based on dither modulation and modification with re-compression

TL;DR

This paper introduces a robust adaptive steganography method that uses dither modulation and re-compression to improve message extraction reliability over lossy channels like JPEG, while enhancing security by embedding in low-frequency DCT coefficients.

Contribution

It proposes a novel modification with re-compression scheme and low-frequency embedding to improve robustness and security in adaptive steganography, outperforming previous methods.

Findings

Achieves strong robustness against JPEG compression.

Significantly improves security compared to prior methods.

Reduces extraction error rate with additional check codes.

Abstract

Traditional adaptive steganography is a technique used for covert communication with high security, but it is invalid in the case of stego images are sent to legal receivers over networks which is lossy, such as JPEG compression of channels. To deal with such problem, robust adaptive steganography is proposed to enable the receiver to extract secret messages from the damaged stego images. Previous works utilize reverse engineering and compression-resistant domain constructing to implement robust adaptive steganography. In this paper, we adopt modification with re-compression scheme to improve the robustness of stego sequences in stego images. To balance security and robustness, we move the embedding domain to the low frequency region of DCT (Discrete Cosine Transform) coefficients to improve the security of robust adaptive steganography. In addition, we add additional check codes to further reduce the average extraction error rate based on the framework of E-DMAS (Enhancing Dither Modulation based robust Adaptive Steganography). Compared with GMAS (Generalized dither Modulation based robust Adaptive Steganography) and E-DMAS, experiment results show that our scheme can achieve strong robustness and improve the security of robust adaptive steganography greatly when the channel quality factor is known.