A Zero-watermarkin Algorithm Using Edge Enhancing

The network is like a double-edged sword, it provides a fast channel for multimedia communication to, but also greatly increases the risk of digital work been copied and modified. Encryption provides an efficient way for copyright protection. The method uses keys and encryption algorithms to change digital work’s information into seemingly irregular clutter information, thus, for those who have no keys and don’t know algorithms, it is impossible to know the protected content. But this will hinder the use and spread of digital work, so we can’t solve these problems only with cryptography. How to protect the legal rights and copyrights of the owner, this is what the digital watermarking will handle.The watermark, as an effect way to protect copyright, needs to be robust, namely, the watermark information can be well extracted after the digital work been processed by conventional image operations. According to the low robust defects on sharpness, dirty diagram of previous zero-watermark algorithms, this thesis puts forward a kind of zero-watermark algorithm, combined with edge enhancement. The algorithm uses singular value and the stability of the gradient to improve the anti-jamming and anti-noise ability.First this algorithm uses the Arnold transformation to scramble the watermark information, and encrypts the treated watermark information with Logistic chaotic sequence. And then the carrier image is processed by block singular value decomposition conversion, and the largest singular value composition of each block is extracted as image thumbnails of the original image. Next, the gradient magnitude is calculated and binarized to get binary image with edge enhancement, which is the feature information of the image. Finally the treated watermark information and image feature information do exclusive or operation; the results will be the key to extract the watermark.The experimental results show that the algorithm can effectively resist the image processing operation attacks, such as filtering, sharpness, histogram equalization, compression, cropping, and has strong robustness.