Protecting data transmitted over the Internet has become a critical issue driven by the progress in data digitalization and communications networking over the past decade. The content being transmitted can be in the form of images, text and voice. To ensure that transmitted data are secure and cannot be tampered with or noticed by malicious attackers, several approaches have been proposed. Steganography is one general approach among them. The hiding capacity and image quality of stego-images are two major measures with which to evaluate an image hiding scheme. To enhance the hiding capacity of Fridrich et al.'s scheme, an improved image hiding scheme for grayscale images based on wet paper coding is proposed in this paper. The significant difference between Fridrich et al.'s scheme and ours is that we shuffle all the pixels in the host image by adopting toral automorphism, then segment all pixels before using our proposed wet pixel hiding strategies to hide the data. Experimental results show that our proposed scheme embeds a larger-sized secret image while maintaining acceptable image quality of the stego-image better than Fridrich et al.'s scheme does. Moreover, the computational cost of our proposed scheme is significantly less than that of Fridrich et al.'s scheme because inverse matrix and multiplication operations are not required. Therefore, our proposed scheme is suitable for real time applications.
IEEE International Symposium on Circuits and Systems, 2009. ISCAS 2009:501-504