Research Of Digital Image Watermarking Attacks And Countermeasures

Digital watermark is a novel and effective method for copyright protection. It becomes a hot spot in multimedia information security research and is a core branch of digital hiding technology. Watermarks were embedded into the digital media and then can be detected to track the used situation of the media. It is useful in information hiding, storing and copyright protecting. At present, watermark technology becomes a kind of effective means to multimedia copyright authentication and its integrity safeguarding. In practical application, the watermarked image will be attacked intentionally or unintentionally, therefore the watermark attack and its countermeasure are the important research direction in watermarking research domain.This dissertation emphasizes on following issues:After analyzing large number of existing watermarking schemes, the digital watermarking technology and basic principle are concluded. Classification and characteristic of the digital watermark as well as to the watermark algorithm evaluation method are summarized. The importance of research of watermark attacks and its countermeasure and the current research situation are analyzed.For further understanding, the existing digital image watermark attack is discussed, the influence of the common several kinds of attack of watermarked image are summed up. At the same, the corresponding countermeasures are given.Base on the study of the existing digital watermark classical algorithm, three kinds of algorithms, which character good anti-attack are proposed respectively. There are bland watermarking algorithm based on relationship of coefficients in DWT domain which can be used to anti cut attack, DCT-based image watermarking using subsampling which can be used to anti JPEG compression attack, and a DWT watermarking algorithm based on singular value decomposition which can be used to anti RST attack. The frame models are proposed and the key problems are discussed for each algorithm, such as, scrambling, the position choice and the determination of the values of the control parameter etc. Lots of simulation examples proved that the three algorithms had good robustness and invisibility.