High-Performance Steganographic Coding Based on Sub-Polarized Channel

Haocheng Fu, Xianfeng Zhao, Xiaolei He

Abstract: Steganographic coding is the core problem of modern steganography under the minimizing distortion model. Syndrome-Trellis Codes, designed with the Viterbi algorithm of convolutional codes, have been the only coding scheme approaching the rate-distortion bound for almost a decade. Although polar codes has shown to have the potential to construct optimal coding schemes, the low diversity of steganographic coding negatively affects the security of steganography, and the performance of the state-of-the-art coding schemes is still unsatisfactory for large-scale applications. This paper proposes a high-performance steganographic coding scheme, named Sub-Polarized Steganographic Codes (SPSC), with near-optimal efficiency and lower computational complexity. By optimal embedding theory and channel polarization, we first establish a universal model of polarized steganographic channels for the coding of steganography. Based on this, the steganographic channels are divided into a combination of multiple sub-polarized channels to construct efficient steganographic codes by updating sequential bit-wise coding to segmented coding. The proposed coding scheme is also evaluated under four polarized channels with typical patterns, of which corresponding sub-coding schemes are also illustrated. Experimental results show that the proposed steganographic coding scheme improves security by increasing the embedding efficiency of steganography, and significantly decreases the computational complexity.