Human Visual System And Its Applications In The Field Of Information Hiding

This thesis discusses the Human Visual System and its applications in the field of information hiding in terms of digital images. First, we analyze the characteristics of image classification methods based on the Human Visual System and then present several image classification principles using HVS for both spatial and frequency treatment and implement a universal image classification software. The software can be used not only in Information hiding but also in other aspects of image processing. Experiment results show its effectiveness in classification. Next, such HVS-based image classification approach is respectively put into use in two important branches of Information hiding, namely, Steganography and digital watermarking and some useful strategies are given accordingly. Algorithms using these strategies are of good robustness, and more importantly, enhance the effect of hiding secret information. Moreover, the thesis also addresses the issue on the implementation of the so-called MPSNR standard (Masked Peak Signal to Noise Ratio) for evaluating image quality, based on which a simple and practical criterion in conformity to Human Visual System is presented. Finally, an Information hiding software system based on HVS is implemented using Matlab tool.