The Research On Digital Watermarking Algorithm Based On Non-subsampled Shearlet Transform

With the rapid development of computer network,there are serious security problems in the transmission of multimedia files.As one of the important means of copyright protection of multimedia files,digital watermarking has been widely used in copyright protection.The Multiscale Geometric Analysis is proposed,which can be used to optimally represent the image or even higher dimensional signal.Shearlet,as the newly developed Multiscale Geometric Analysis tool,has more sparse expression ability and stronger directional sensitivity.Non-subsampled Shearlet Transform not only reduces the down-sampling operation in the standard Shearlet transform,but also overcomes the frequency spectrum aliasing phenomenon and improves the directivity.Based on the study of Shearlet theory,the Non-subsampled Shearlet Transform is applied to the field of digital watermarking,using its excellent characteristics,such as sensitivity to the direction,the most sparse expression ability,etc.,to further improve the performance of image watermarking.After the image is decomposed by the Non-subsampled Shearlet Transform,the most significant directional subband image is selected as the embedding position of the watermark.Then,two efficient algorithms are used to embed watermark.The watermarking technology is applied to the identification of the authenticity of electronic documents.The main content of this paper are as follows:A non-blind watermarking algorithm is proposed based on the combination of the Non-subsampled Shearlet Transform(NSST)and SingularIV Value Decomposition(SVD).The image is decomposed into the low-frequencyapproximate subgraph and the high-frequency detail subgraph after the Shearlet Transform.In the experiments,selecting the directional detail subband image with the maximum energy as the embedding subband of the watermark,and adding the scrambled and encrypted watermark information to the singular value matrix of the subband image by using Singular Value Decomposition,watermark is embedded.Adding a small perturbation to the image,the size of the singular value is basically unchanged.This stability is helpful to improve the watermark performance.Different types of attacks are used to do the image which embed the watermark,and then the watermark is extracted,proving the watermark's resistance to external attacks.An image zero watermarking algorithm is proposed based on the Non-subsampled Shearlet Transform and Principal Component Analysis.After the image is transformed by Shearlet Transform and the watermark is embedded in the subband,while the Principal Component Analysis is used to reduce the dimension of the subband image,and obtains the important information of the image.Using these image features,combined with the watermark information construct the zero watermark,so as to achieve the watermark encryption.Zero watermarking technology is not the true meaning of embedding,so there is no damage to the image quality.Zero watermarking is a kind of digital watermarking technology which can not modify the original image data.It can use the important features of the image to construct watermark information,instead of modifying these features of the image.In addition,the main features of the image obtaining by the Principal Component Analysis transform,Which include the low frequency components of the image,and the high frequency details such as texture,edge and so on,can effectively resist different forms of attack.In this paper,efficient algorithms are proposed,and the watermark embedding does not distort the original image,which indicates that thewatermark has good invisibility.In the experiment,after the watermarked image is attacked by different forms,and the watermark can be extracted very well,which it shows strong robustness.There are many excellent characteristics of the Non-subsampled Shearlet,such as the best approximation to the image and the sensitivity to the image.The combination of these excellent features and watermark making the two algorithm proposed in this paper truly achieve the invisibility and robustness of watermark.The application of digital watermarking technology in the electronic certificate to identify the authenticity of are analyzed.In order to improve the security performance of the electronic certificate,the personal information is generated into the QR bar code,to be embedded as a watermark into the document,and through the application of efficient algorithm is proposed.The experiments show that the algorithm is efficient and robust,the QR code can be extracted very well and can be decoded successfully.To achieve the purpose of electronic certificate anti-counterfeiting identification,demonstrates the practicability in practical application.