Research And Application Of Digital Wartermarking Algorithm Based On DCT Transform Domain

With the rapid development of Internet and the rapid process of digital media resources. The copyright of digital product(such as images, audio, video, etc.) problem is increasingly highlighted, and makes the digital watermark technology become an important subject in recent years. Especially in the field of information security, computer science, communication field has caused researchers’ extensive concern. Since the 1990’s, some high levels of academic journals are published papers about digital watermark, and the number of articles also increased year by year. Then some international conference and magazines also published a special issue about digital watermarking. Foreign universities and research institutions are also committed to the research of watermarking algorithm, such as the United States MIT University,the Purdue University and the England Cambridge University. The institute of NEC research in Japan and the institute of IBM in the United States, etc. Through the unremitting efforts have achieved certain scientific research achievements.This thesis is based on the research of the DCT domain watermarking algorithm for digital image. First of all, the author overviewed the information hiding technology and the development of digital watermarking, and then analyzed the characteristics of digital watermarking, application field, classification standards, and the basic theory and based on DCT domain digital watermarking algorithm----one of the mainstreams of the current algorithm. Later the author introduced scrambling algorithm and chaotic encryption algorithm used for watermark preprocessing and applied to the digital watermarking technology systematically. Finally, digital watermarking algorithm is realized based on Arnold scrambling and chaotic encryption method to strengthen the security of digital watermarking algorithm.In this thesis, the author implement two kinds of digital watermarking algorithm based on DCT transform domain. The main idea of the first watermark algorithm firstly is blocking the embedded watermark image firstly, and then transforms each sub-block by DCT, and to Arnold scrambling watermark image information and chaos encryption, finally select coefficient in of each sub-block image DCT. And realize the watermark embedding according to certain algorithm. The second one, proposed on the basis of the improvement ideas, is a method that realizes the watermark embedded using the replacement of intermediate frequency coefficient. The innovation of this algorithm is the proposed is not directly into 1 or 0 superimposed on watermark embedding coefficient, but with two 1 row 8 column of random sequence on behalf of the watermark information 1 or 0, the watermark embedding coefficient overlay said random sequence, and add the two adjustable parameters, which can respectively adjust the intensity and position of the watermark information. This method has the advantage that not only realizes the blind extraction of the watermarking algorithm, but also greatly increases the accuracy of watermark detection.In this thesis, the author mainly studied the DCT domain watermarking algorithm, designed and implemented two types of watermark embedding algorithm, and checked the property of the two kinds of watermarking algorithm through a lot of simulation experiments. Generally speaking, the two kinds of watermarking algorithm show the strong robustness in resistance to a variety of attack types. The two kinds of watermarking methods needn’t original image in watermark detection both, fully realize a blind extraction. The second kind of watermarking algorithm is in summarizing the advantages and disadvantages of the first kind of watermarking algorithm on the basis of the author to do the further research, so the results also show the better performance against attacks such as JPEG compression, noise, filtering in the second algorithm. In this paper, a kind of algorithm of improvement is shown for the both watermarking method lacking of respect in the resistance to shear attack. The experimental results show that the algorithm improved is robust for resistance of shearing attack is put forwarded, and the design achieved the expected effects. At the same time, the robustness of the algorithm is not lower in the other respects beyond resistance to shear attack.