The Contents Characteristic Based Digital Image Watermarking Method Research

The growth of the digital multimedia technology and the successful development of the Internet, various forms of digital works such as image appear in large numbers. The problem of intellectual property rights has caused wide concern. However, digital watermarking is an effective means of copyright protection. With no impact on the digital works premise, the digital watermarking embeds watermarking information into the digital data. When appears copyright disputes, through the watermark detection and analysis on its works to proof the creator. Therefore, the feature of the high security, invisibility and robustness are necessary for the scheme of digital watermarking and the contradiction between the invisibility and robustness also need be considered. In this paper, in order to balance this contradiction and resist the synchronization attack, two kinds of image watermarking algorithms are present:1, Human Visual System based digital image watermarking has been presented. First of all, a significant region is procured from the geometrically invariant space that is constructed through image normalization. Secondly, according to Human Visual System (HVS) masking, we calculate the significant region of visual masking. Thirdly, to embed the watermark into the DFT domain, the conceal quality of HVS is used. Experimental results show that the proposed scheme is not only invisible, but also robust against common signals processing and geometric attacks.2, the Bayesian image segmentation based digital image watermarking has been presented. First of all, the Bayesian image segmentation is used to segment the cover image into several homogeneous regions. Secondly, the method of combing the feature scale theory to confirm the local feature regions as well as the extraction of the feature points using the Harris-Laplace detector is adopted. Thirdly, in terms of the conceal quality of HVS, a watermark with readable characteristic is embedded the DFT domain. The simulations prove the robust characteristic of the proposed method is not only feature points distributed equably but also robust against common signals processing and desynchronization attacks.