The Research On Capacity Of Digital Image Watermarking

With the rapid development of information technology and computer network, the storage, duplication and propagation of digital multi-media information has become more and more convenient, and the pirate problem and copyright dispute initiate and become worse gradually. The watermarking technology is to come into being under such background. It embeds special information into digital medium, and the information can be recovered when needing, and this can reach the purpose of protecting copyright.In this paper still digital image watermarking is selected as research object. the background of digital watermarking and the relationship between digital watermarking and information hiding is introduced; the development status of digital watermarking is outlined and some typical image watermarking algorithms are studied in detail, classification and common applications of image watermarking digital watermarking are summarized; the classification of the watermarking channel is made and the current capacity theories are analyzed; the human visual system is introduced and its characteristics are analyzed in detail; Image watermarking capacity is computed and analyzed based on human visual model. The main content is as follows:1. Watermarking channel model is introduced and the current capacity theories is analyzed in detailSome commonly used models of digital watermarking, including Costa model, Moulin model, Cohen and Laplidoth model, parallel Gaussian channel model and so on, are studied in detail; Their capacities are described respectively; some conclusion is drawed that these models are based on different assumptions: different model, different assumptions will lead to different results and capacity formulas.2. The human visual system'characteristics are introduced and analyzedThe characteristics of human visual system, including the brightness, contrast sensitivity and texture hidden characteristics, are analyzed in detail, several common models of the human visual system are described. 3. Image watermarking capacity is calculated based on the human visual modelThe spatial domain capacity is calculated and analyzed based on Chou and Li model, and the corresponding changes relationship between the capacity and peak signal to noise ratio is given; then the DCT domain watermarking capacity is studied based on Watson model, and an adjustment factor is set up. the relationship graphs between adjustment factor and capacity, peak signal to noise ratio are obtained; Finally the DWT capacity is studied based on the noise visibility function.4. Capacity comparison and analysis of different models in the different domainsFirst the capacity comparison in different domains is performed, including spatial domain, DCT domain and wavelet domain; then the capacities of Chou and Li model, NVF model in the spatial domain are compared; then the capacity comparison between spatial domain and wavelet domain based on NVF model and the capacity comparison between DCT domain and wavelet domain based on Watson model are performed;finally the comparison capacity based on information theory and JND model is performed.