Research Of Image Watermarking Algorithm Based On Discrete Cosine Transform

The rapid growth of the IT(Information Technology) makes electronic media communication easier and more extensive than before . While this enables the illegal creation and dissemination of products quickly via electronic means . There have emerged many techniques for protecting the security of multimedia information in recent years . Image digital watermark is a booming technique , which, in fact, is information hiding technique using in image processing field . The key to design a good watermarking algorithm is invisibility and robustness.First, some basic knowledge of digital watermarking techniques are systematically discussed, including its background, general system structure, basic principle, state-of-the-art, and their application. Secondly, as an important branch of information hiding, we analyzed the current watermark developing status, after deeply analyzing the technical characters and request of image watermarking, this paper proposes a new blind watermarking algorithm that is based on classifying the image blocks and uses the spread spectrum technique and two keys to embed watermark in the image, and the watermark is decoded or detected without the original images. The proposed algorithm is different form common method in that, for the latter, a coefficient is embedded a watermark information, and watermark casting is performed by some geometrical attack. We embed one watermark many times in an image, group some selected image pixels or transform coefficients to form a vector in a big block and call it the watermark host vector, and divide the host vector into subvectors of equal length. Each subvector is used to embed one bit of watermark information, so, embed the same watermark information in different blocks. For a good watermark algorithm 'invisibility and robustness, DC, mid_frequency and parts of low frequency coefficients in DCT are selected as location where to embedded a watermarking. Finally, on the basis of statistical analysis, two formulas are presented. One is used to calculate the correlation between the DCT coefficients of image to be detected and the inserted watermark. The other is used to calculate the threshold of the correlation. The method of inserting watermark in color image is discussed on the basis of grayscale image.In order to test the imperceptibility and robustness of of the algorithms, simulation experiments are conducted. The experimental results indicate that: the new blind watermarking algorithm is imperceptible and robust to common attacks such as JPEG compression, low-pass filtering, noise contamination, cropping, zooming, multiple watermarking, collusion attack etc.. The research of this paper is occupying a forward position in relevant field. It has important theoretical value and application prospect. The proposed methods and experiment results in this paper have important reference value both in further studying of digital watermarking and in developing relevant computer software.