Font Size: a A A

Research On Invisible And Robust Watermarking Algorithm

Posted on:2019-11-28Degree:MasterType:Thesis
Country:ChinaCandidate:T GaoFull Text:PDF
GTID:2428330572452538Subject:Software engineering
Abstract/Summary:PDF Full Text Request
In order to coordinate the contradiction between the invisibility and robustness of watermarking algorithm and enhance the ability of resisting geometric attack,an invisible and robust watermarking algorithm based on image blocks is proposed.The algorithm firstly divides carrier image into equal-sized and non-overlapping image blocks,the masking value of the image block is obtained by the texture information and edge information,and the image block with a higher masking value is selected as an embedded sub-block.Then,the selected embedded sub-block is subjected to discrete wavelet transform(DWT)and the resulting low-frequency subband is decomposed by singular value decomposition(SVD).According to the watermark bit information,the pair of elements with the largest difference in the first column of the orthogonal U matrix is selected as the embedded position,and the Arnold scrambled watermark information is embedded.Before detecting the watermark,the watermarked image and the preserved template are matched with the feature points of the scale invariant feature transform(SIFT)to determine whether the watermarked image is subjected to geometric attacks.If geometrically attacked,geometric correction of the watermarked image with geometric distortion is realized by coordinates and scale information of the SIFT feature points,which restores the synchronization of the watermark and the carrier image.Experiments on standard gray scale images,the peak signal-to-noise ratios(PSNRs)of the watermarked image can reach more than 44 dB,which shows that the image masking analysis can be effectively carried out by using the texture and edge features of the image block,thus ensuring the invisibility of the watermark algorithm.When carrying out common attacks and geometric attacks on watermarked images,the SIFT feature points are used to correct the watermarked images after the attack.The normalized correlation(NC)coefficients between the extracted watermark and the original watermark can mostly exceed 0.99,indicating that the algorithm has strong robustness to the common attack and the geometric attack.The experimental results show that the algorithm can effectively resist geometric attacks and better coordinate the contradiction between invisibility and robustness.
Keywords/Search Tags:invisibility, robustness, discrete wavelet transform(DWT), singular value decomposition(SVD), scale invariant feature transform(SIFT), geometric correction
PDF Full Text Request
Related items