Semi-fragile Watermarking Algorithm For Image Tampering Localization

Along with the extensive using of the web services, the technology of authentication hasaroused more and more concern. At present, digit watermark as an effective measure ofresisting image pirating and distorting has greatly applied. In order to protect the interest ofindividual and social from damage, and guarantee the order of internet efficiently, we needpropose an algorithm and technology to guarantee the authenticity of digit image efficiently.In order to reliably detect the integrity of the watermarked image, in this paper, wepropose two algorithms which are used for resisting image distorting. One is based on thetechnology of lattice vector quantization, the main idea is that first we divide the host imageinto three parts randomly, then generate the watermark information according to logisticmapping function, and embed the watermark into the blocks of three different parts of the hostimage, in which any block of a part can be the positioning reference block of other particularblocks in the other two parts. The use of the technology of lattice vector quantizationguarantees the quality of the watermarked image preferably, and improves the locationaccuracy effectively. The other semi-fragile watermark algorithm is combined the DCTtransform in the wavelet domain. Before embed the watermark, we use Logistic Chaos reflectto encrypt the original watermark image; in the process of the watermark embedding, we firstimplement the sub-block DCT transform in the host image wavelet domain, then according tothe image content features, we embed the watermark into the DCT coefficient matrixself-adaptation. In the image tampering positioning process, the original host image is notused, making the operation more convenient and flexible.Through many experiments, the results indicate that, the first algorithm proposed in thispaper can endure the incidental distort at a certain extent and can position the tempering areaswith high accuracy. To further improve accuracy of locating and realize recovering of thedistorting regional content are the future work. The second algorithm we mentioned in thispaper can guarantee the quality of the image, on the basis of that it also can position themodify location of the image accurately. In the experiment of JPEG compress, noise and otherattack, the algorithm show strong robustness, and the all above satisfy the generalrequirements.Semi-fragile digit watermark algorithm is the main realization method of the digit imagedistorting location system in this paper. The each stage of the modeling process adopts thetools, such as Visio Studio2003. In program development stage, as the Visual Studio2008issimple to use and has better expandability, we choose it to be the programming environment of this system. And the API function used for digit image processing integration in theOpenCV1.0can ensure the program development of high efficiency.Generally speaking, we proposed two kind of semi-fragile digit watermark algorithm.Compare with the algorithm existed before, we do the improvement in the authenticity andintegrality based on the image content. The ability of distort location and detection improvedremarkable. The improved algorithm overcomes the disadvantage of original algorithm, aswell as makes a good effect and user experience, so that is no doubt that the improvedalgorithm provide convenient and guarantee for each kind of people in the society.