A Compound Watermarking Algorithm For Digital Rights Management

With the rapid development and comprehensive application of internet and multi-media communication technology, the problem of infringement and piracy of digital products is becoming increasingly serious, copyright protection of that has become an urgent problem need to be resolved, digital image watermarking technology is an effective solution to solve the problem. Digital watermarking technology makes signs of copyright information such as serial number, text, images and other information embedded in multimedia data in order to play a role of copyright-protected, secret communications, authenticity and products identification. Digital watermarking includes the following three characteristics:Concealment: digital watermarking can not be perceived to be, that is, the existence of the watermark should not significantly interfere with protected data and affect the normal use of protected data.Robustness:the watermark must be difficult to (cannot would be the best) be removed, if we are only aware of part of the digital watermark information, then any attempt to remove or destroy the watermark will lead to serious quality drop and make that is not available.Security: Information in the digital watermarking should be safe, meanwhile it should be difficult to be tampered with or forged; it should also have a lower rate of false detection.According to the hide location of digital watermark, watermark characteristics, media attachment, watermark purposes, detection process, content, watermarking can be divided into a variety of categories:grouped by the characteristics, watermarking can be divided into robust and fragile digital watermarking; grouped by the media attachment it can be classified into image watermarking, audio watermarking, video watermarking, text watermarking, as well as grid watermarking for three-dimensional grid model and so on; grouped by the detection process it can be divided into clear watermarking and blind watermarking expressly; grouped by the content it can be divided into meaningful and meaningless watermarking; grouped by the usage it can be divided into notes security watermarking, copyright protection watermarking, tampered remind and hidden logo watermarking; grouped by the hidden location of the watermark, it can be divided into time (space) domain digital watermarking, frequency domain digital watermarking, time/ frequency domain digital watermarking and time/ scale domain digital watermarking. With the development of watermarking technology, endless variety of watermarking algorithms are emerging, it should be mentioned that, as long as the signal transformation composition is possible, it is possible to hide the watermark in its transformation space.Of course, as a means of copyright protection, digital watermarking is bound to encounter various forms of attack. Strong attack on the watermark is an intuitive evaluation approach for the robustness of the. watermark. It is like the negation of negation process helical improvement of traditional encoding-decoding. There is a need to research watermark attack thus we can have a better understanding and discovery of the existing weaknesses in the watermark embedding technology, and it can also promote the watermarking technology stronger and better.From the principle that the mode of attack to the digital watermarking can be divided into: de-noising losing compression, re-modulation attacks, average conspiracy attacks, geometric attacks, cryptographic attacks, protocol attacks and the second generation attacks, as well as quadratic attacks, etc. However, in most cases the attacker is not using only one means of attacking, most of the time they adopt a comprehensive approach,In order to cope with the watermarking attack efficiently, in my opinion, a good digital watermarking scheme must meet the following three demands:first of all, the embedded watermark must possess good robustness and transparency, thus it can resist malignantly destructive attacks. Secondly, we should embed some more watermarks if the capacity of embedded watermark permits. It can not only resist various kinds of attack, but also misdirect the attacker, make the attacker take it for granted that he has won drastically when he breached the first watermark. Thirdly, also the most important, it is that the embedded watermark must be proved to be owned by yourself. This is very important, if you can not provide enough proof, all the work you have done would make no sense.In view of the last three points, an algorithm of compound image watermarking based on wavelet transformation was proposed in this thesis. Its theory is that, the logo copyright information image and author information are embedded in loaded image using various kind of mode, when the watermarks are extracted, using logo copyright information, author information and the date embedded to prove the ownership of watermark, accordingly the copyright is protected effectively.The reason why choosing wavelet transformation for analyzing is that wavelet analysis has a good performance in signal and image processing. It can be used in the handle of boundary and filtering, time-frequency analysis, signal-noise separation and extracting weak signal, solving fractal index, signal recognition and diagnose as well as multiple-scale edge detection, etc. In view of image compression, it can do better than other method not only that it can make the feature of signal and image persistent after compressing, but also that it can resist disturb when passed and the compression ratio and speed are much better.The detailed information of the algorithm is that: firstly, embedding copyright mark image information, and embedding the frequency domain WLL2, WLH2, WHL2, WHH2 of copyright mark sequence respectively into the frequency domain IHH4, IHL4, ILH4, ILL4 of carrier image according to the strength factor of the wavelet after copyright mark image and carrier image are transformed by wavelet.Afterwards, embed author information image, make the carrier image with copyright mark information be transformed by third level wavelet, lower the dimension of the author information and encrypt it, meanwhile, embed it in the frequency domain ILH3, IHL3 of the carrier image embedded in the copyright mark information using spread spectrum technology. The encryption key is designed based on copyright information and embedded date of watermarking.As extracting watermarking, extract author mark image according to encryption key utilizing wavelet inverse transformation, after twice wavelet inverse transformation, extract copyright mark image,In this way, the ownership of watermarking can be proved by copyright mark, author label and embedded date of watermark.The experimental result indicates that the watermarking proposed in the thesis is robust and covertly, it can also resist various kind of attack like second attack, image cutting, compressing, noise attack, etc. But it also needed to be improved in terms of rotating attack, filtering attack and so on.