Studies Of Color Image Digital Watermark Algorithm

With the rapid development of Internet technology, the spread speed of digital information becomes faster, and traditional information is digitalized. Digital information has the advantage of communication convenience inherently. However, this advantage also requires more in the copyright protection of information. The copyright protection and falsification preventing of digital products are hot topics in the Internet era. It has been deeply researched how to use technology to protect the copyright of digital products, while reserving their communication convenience. Digital watermarking technology is developed to solve this problem. Digital watermarking can be used in digital products (e.g. images, audios, videos, etc) to protect their copyrights.First, this paper introduces the principle of digital watermarking, including its basic feature and the research background in the academic area. In the mean time, this paper introduces the basic framework of digital watermarking, and states the all components of this framework. And then, this paper classifies the current digital watermarking algorithms according to their features, which provides the theoretic basis for the design and implementation of digital watermarking algorithm. The DCT as well as the watermarking algorithm based on it are deeply analyzed. Because the watermarking algorithms base on DCT are widely used and own great robustness.After that, work for color digital image watermark algorithm is introduced. It makes use of the features of color images, and presents the concept of anchor point in Lab color space. The algorithm proposed in this paper is robust against both geometric and non-geometric attacks. It overcomes the shortcomings of lacking of robustness under non-geometric attacks, which exists in most algorithms present. And at the same time this algorithm do little damage to images, so it is more powerful in practical use. Besides that, this paper also improves an algorithm based on spatial domain to frequency domain. Experiment results show that the algorithm proposed in this paper has stronger robustness, particularly in the geometric attacks. This algorithm has already been used in a copyright protection project.At last, a conclusion is drawn and future work is presented.