Hidden content quality aware stego-image hiding method using re-encoding strategy

摘要

In normal Steganography, the main motivation is to maintain high quality of the stego image, so that one is not suspicious that there is a hidden image. At times, it is important that the hidden image quality needs to be maintained after recovery. How to achieve high quality of hidden image is the motivation of this work. To solve this problem, before the hiding procedure, pixels of the secret image are analyzed to generate the optimum code book. Most frequently occurring pixel value is encoded with the shortest code to minimize stego-image distortion. For evaluation, we not only tested the distortion of simple hidden images like a logo, but also high resolution images. Using the proposed method, PSNR value of more than 45.6 db for the stedo-image was achieved even for high resolution hidden image. Secret image PSNR after recovery was 50 db or more. We could conclude that the proposed method can yield good results regardless of the type of hidden image.