Digital Watermarking Algorithm Research On Beyond Wavelet Domain

The Popularity of the Internet has shortened the distance between everyone, itallows us to be able to enjoy a more convenient and faster real-time communication.But the resulting problems constantly plagued people. The original author obtainedthe results after their efforts be copied by unlawful elements and keep it for himself.That phenomenon has seriously damaged the interests of the creator. As an effectivemeans of protection to multimedia data, digital watermark is simple, efficient withits unique advantages both at home and abroad in recent years has aroused greatinterest of people.Wavelet theory is the hot technology at academic in recent years, because of thegood time-frequency analysis capabilities, it has been become one of the mostimportant technical in engineering. However, some of the inherent flaws of thewavelet has affected its be used in the broader field. So, this paper embarks from theoptimal representation of signal, analyze and verify the excellent performance ofbeyond wavelet. according to different characteristics of beyond wavelet coefficients,we has researched the application of digital watermarking in beyond waveletdomain.In this paper,according to the different original carrier,we have proposedwatermarking algorithm based image and video.Specific research contains thefollowing three:(1)Wavelet decomposition of the digital image can not be sparse expressed. andthis will affect the efficiency of the wavelet coefficients, reducing the strength of theembedded watermark, thus affecting the robustness of the watermark. we haveaccording to contourlet decomposition coefficients characteristics, design andachieve a robust watermarking algorithm based on the neighborhood means relationsin contourlet transform. Firstly, watermark image scrambling encryption with arnoldtransform. Then do Contourlet decomposition of original carrier, achieve watermarkembedding with the relations between coefficients of low frequency decomposition.It’s not required for original image in the watermark extraction, we have achievedthe blind extraction. (2) Tamper detection and recovery capability has been a fatal flaw of fragilewatermarking in the transform domain.Therefore,this paper presents aself-embedding fragile watermarking algorithm based on DDWT image coding.Tamper localization of the algorithm and tamper recovery of image are respectivelycorresponding to the authentication watermark and recover watermark. The SPIHTcoding compression image of the original image in the DDWT is the recoverywatermarking, in order to enhance the robustness of the watermarking, we hadembedded it into low frequency of wavelet decomposition. The Authenticationwatermarking is the characteristics of the relationship between low-frequencycoefficients, and its embedded in the image sub-block of LSB.(3) For the current, people are increasing interested in video streaming, this paperpresents a video watermarking algorithm based on optical flow field motionestimation. Firstly, detected the key frames of the video by optical flow information,Then, carry out motion estimation of the local region in key frames by optical flowfield. according to the sensitivity characteristics of human eye to the rapid flowobject, the watermarking is sub-blocks embedded in the region of rapid movement inthe key frame.