Modified Matrix Encoding Based on the Spatial Distortion Model and Its Improvement

Steganographic coding is a common and effective method to improve the security of steganography. In the meantime, how to define a suitable distortion model for different cover objects is another crucial problem in the design of steganography. In this paper, we define a new and simple distortion model for spatial images and apply it to the MME (modified matrix encoding). It is used to choose the pixels whose modifications introduce the minimal embedding distortion. In addition, an efficient and simple steganographic code, based on the LSBMR (least significant bit matching revisited), is proposed to further decrease the changes of the cover elements, which is used to improve the MME based on the spatial distortion model. At the same time, the idea of the proposed method can be similarly extended to the STC (syndrome-trellis codes) and the widespread block codes based matrix embedding schemes such as BCH (Bose-Chaudhuri-Hochquenghem) codes. The experimental results demonstrate the rationality of the proposed distortion model for spatial images and the efficiency of the proposed steganographic code. Moreover, when resisting some common and efficient steganalyzers, the proposed steganographic algorithms, tested on 10 000 grayscale images of BossBase1.01 image library, perform better than some previous papers.