Research On Reversible Digital Watermarking Algorithm By Prediction Error And Image Interpolation

With the continuous development and progress of Internet,people enjoy the convenience of modern information technology,and also face the hidden dangers of information security.Some illegal criminals illegally tampered with large amounts of digital products through improper ways,and then tampered with them and propagated their own interests,which seriously damaged the rights and interests of the copyright owners of these products,and caused a lot of security problems to the information industry.Digital watermarking can solve this problem,watermarking embedded into digital products with special significance,in the copyright problem cannot be authenticated,the watermark information can be extracted by products with special mark,which proves that the product copyright,to provide help for this kind of dispute.However,the traditional digital watermarking technology will bring irreversible damage to the carrier when it is extracted.It is not suitable for application to the field of high requirement for the quality of the original carrier.In addition,the reversible digital watermarking technology can completely detect the watermark information and extract the original vector without distortion.Because of the addition of the feature,reversible watermarking technology has a particularly important application in special fields.In this paper,grayscale image is used as an experimental carrier,and the existing algorithms are improved from two aspects of image imperceptible and watermark embedding capacity.The following are the main contents and innovations of the paper:(1)This paper briefly describes the research background and development process of digital watermarking technology,expounds the basic knowledge and related algorithm of reversible digital watermarking technology,and provides a theoretical basis for the algorithm design in this paper.(2)In order to improve the performance of the reversible image watermarking algorithm,this article proposes a reversible image watermarking algorithm based on second-prediction-error and pixel sorting.Firstly,the second-error of pixels is predicted,and the prediction error of the remaining pixels is determined by using the error of adjacent pixels.Then,the pixels are sorted by local complexity so that the pixels with the smaller error value are first processed,and the data is embedded into them.Finally,the zero gap is generated by shifting difference histogram,and the watermarking information is embedded in the zero gap of redundancy.The experimental results show that the prediction accuracy and the image distortion can be improved effectively by usingthe second-prediction-error and the pixel sorting technology,compared with the related algorithms,PSNR of the watermarked image is improved by about 1dB in the same embedding capacity,and the algorithm has better imperceptibility and performance.(3)In order to solve the problem of improving the watermark embedding capacity and causing the image inability to decrease,a reversible image watermarking algorithm with direction interpolation and differential expansion is proposed.The algorithm combines the difference extension,the histogram shift and the image interpolation to complete the embedding of the watermark.The algorithm first classifies the pixels of the original image,and then interpolates and forecasts the pixels after classification.The prediction value is different from the original pixels,and the calculated difference is expanded and embedded into the watermark.The pixels that cannot be embedded with watermark are translated by histogram statistics.In particular,the algorithm is used as a reference pixel for prediction,and it can also be applied to watermark embedding,improve embedding capacity and image quality.