Research On Technology Of Semi-fragile Watermarking For Image Authentication

Digital watermarking technique is a whole new mean to guard information security. It has been a strong method for copyright protection and authentication on multimedia. And it involves the theory on communication and information, digital image processing, network security and so on. It is of high academic value and economic value.Semi-fragile watermarking technique is an important branch of digital watermarking. It can tolerate general signal processing, detect any spiteful tamper on marked multimedia data, locate spatial regions within an altered image, and even can estimate what kind of modification had occurred and recover the tampered data. With these attractive characteristics, semi-fragile watermarking has become a research hotspot of multimedia authentication domain in recent years.Discrete Cosine Transform and Wavelets are respectively the key technique in the subsistent source compression coding standard JPEG and source compression coding standard JPEG-2000, the background, the concept, the general framework and the research trend of authentication watermarking (known as fragile watermarking and semi-fragile watermarking) technology is roughly described, the emphasis is on the analysis of authentication watermarking classified by the spatial and the transform domain, also their advantages and disadvantages are analyzed, in addition, common attacks applied to authentication watermarking are summaried.Two novel semi-fragile watermarking algorithms respectively based on JPEG and JPEG2000 are proposed in this paper. Both watermarks are generated based on the relative signs and magnitudes of transform-domain coefficients which are invariant due to an important property of JPEG and JPEG2000 compression. Combined with an effective method to include all the coefficients either explicitly or implicitly, the proposed algorithm improves its security level over malicious attacks in transform domain, without embedding more watermarking information in the original image. With the the former algorithem's advantages of invisibility, block-wise dependence, accurate localization of tampering and easy to be realized, additionally the latter algorithem can prevent the attack of re-embeding false watermark to confuse the authentication of copyright. It is realized by using private chaos key to encrypt the watermarkable points. The experimental results are provided to demonstrate that the algorithems are easy to be realized, accurate to localize the tampering, credible and practical.