Distortion Function for JPEG Steganography Based on Image Texture and Correlation in DCT Domain

Zichi Wang,Zhaoxia Yin,Xinpeng Zhang

DOI: https://doi.org/10.1080/02564602.2017.1304289

2017-01-01

Abstract:This paper proposes a novel distortion function for JPEG steganography, which depends on the magnitude of discrete cosine transformation (DCT) coefficients, the first- and second-order residuals, and the amount of zeros in a DCT block. The magnitude of a DCT coefficient indicates the ability of the current coefficient to conceal modification trace. The first- and second-order residuals make sufficient use of the correlation in DCT domain, and the amount of zeros in a DCT block illustrates the block texture well. These elements are combined together to measure the risks of detection due to the modification on the cover data. By using this proposed distortion function, the embedding changes are limited in texture regions when employing the syndrome trellis coding to embed secret data. Thus less detectable artefacts could be found in stego images. The experimental results demonstrate that the proposed scheme performs less statistical detectability when comparing to current state-of-the-art steganographic methods for JPEG images.