Watermarking Algorithm Against Geometric Attack Based On LS-SVM And QR Decomposition

With the rapid development of Internet technology,the amount of digital information has exploded geometrically.Issues of digital information security and copyright need to be resolved.The emergence of digital watermarking technology provides an effective way to solve these problems.Digital image watermarking is one of the digital watermarking technologies.At present,there are many algorithms for digital image watermarking.But so far,algorithms that can effectively resist all attack methods have not appeared.In these attacks aiming at watermarking algorithms,geometric attacks are a more difficult method to resist are also an important restricting factor of the development of digital image watermarking techniques.Geometry attacks generally do not destroy the embedded watermark information,but they change the position of the image pixels in the spatial domain,making the embedding of the watermark unsynchronized and thus unable to accurately extract the embedded watermark information.After the image is attacked by geometry,the watermark still exists.As long as the synchronization of the embedded watermark is restored,the correct watermark information can be extracted.Based on this idea,this paper designs a digital image watermarking algorithm that can effectively resist geometric attacks.The algorithm idea is as follows:1.Use watermark embedding algorithm based on QR decomposition to embed the watermark in the QR domain of the image.2.The rule moment of image after different RST(rotation,scaling,translation)parameter attacks is calculated.Then the training set is composed of these rule moments and the parameters of the RST attack.The training set is trained using a Least Squares Support Vector Machine(LS-SVM).Get a trained LS-SVM model,then input the moment value of the image that is attacked by the geometry as the model.The output of the model is the estimated geometric attack parameters.The corrected image is corrected once based on these parameters.3.On the basis of step 2,use the template correction method to perform a second recalibration.4.A watermark extraction algorithm based on QR decomposition is used to extract the watermark in the QR domain of the image.Finally,through experimental simulation,the algorithm proposed in this paper has good robustness against ordinary attacks or geometric attacks.