Research Of Resisting Geometric Attacks Image Watermarking Based On Template Matching

In recent years, with the rapid development of the Internet and multimedia technologies, all kinds of multimedia works can be published or downloaded through the Internet. This brings great conveniences to our working and daily lives, but there also comes along with the problem of intellectual property protection. Thus the safety of multimedia product becomes a very important and challenging research. Digital watermarking technology, which is an important branch of information hiding technology, is an effective way for the realization of multimedia copyright protection and the guarantee of the integrity of information. And it is also becoming a popular research in the information field. It embeds secret information called watermark in the raw data to confirm the ownership of the data. So the watermark becomes an inalienable part of the source data, which can experience some operation that do not undermine the use-value or the commercial value of the original data to survive. As geometric attacks greatly affect the robustness of watermark systems, the research on anti-geometric transform attacks is considered as a difficult and tough issue to improve the robustness of watermarking algorithm.This paper is focused on static images only. By utilizing the theories of image processing and characters of watermarking, an affine resistant digital image watermarking based on template matching was proposed. First, the message to embed is coded using RS codes. Then M-sequence is used for spread spectrum to enhance the resistance capacity of noise. Watermark is embedded on the mid-frequency spectrum, and the positions are considered in a rectangle which is symmetric to one side of the cross. While the specific embed positions are determined by a one-way hash function. Two structural template lines are then added in the polar coordinate system for detecting any combination of the geometrical distortions. The detection and extraction is based on correlation between the embedded and recovered watermarks. The main contributions of this paper are as follows:①RS error-correcting coding is used before embedding the watermark, which enhances the error-correcting capability of the watermark in the extraction phase.②Watermark is embedded on the mid-frequency spectrum, and the positions are considered in a rectangle which is symmetric to one side of the Cross. This operation improves the rate of successful extraction of the watermark. ③Re-structure the embedded template formula. Furthermore, local standard deviation and global standard deviation are introduced. In this way, we will get better local peak, and decrease the possible local peak numbers in the extraction phase of the watermark. The Experiment confirms that the improved formula brings a better performance for the system.Finally some drawbacks of our work are also pointed out.