Research Of Zero-watermarking Algorithm Based On Boost Normed Singular Value Decomposition

A novel zero-watermarking algorithm based on Boost Normed Singular Value Decomposition(BN-SVD)is proposed to solve the problem of diagonal distortion and false positives,which are caused by Singular Value Decomposition(SVD)applied to zero-watermarking.First,the watermarking image is disposed with Arnold transformation and Logistic map to obtain an encrypted and scrambling watermarking image.Then,the original image is pre-processed by Discrete Wavelet Transform(DWT)and Discrete Cosine Transform(DCT),and the low-frequency coefficient matrix is obtained.SVD is used to low-frequency coefficient matrix,at the same time,the diagonal matrix is adjusted by using a parameter to achieve BN-SVD.The characteristic vector is calculated by the maximum singular values and the maximum singular values from BN-SVD;Finally,the characteristic vector and the scrambling encrypted watermarking image are both sent into a Cellular Neural Network(CNN).The output image is the registration image of zero-watermarking.The performance of the watermarking algorithm is evaluated by the normalized correlation coefficient.The experimental results show that under the various attacks such as JPEG compression,noise,filtering,rotation and shearing,the normalized correlation value of the extracted watermarking and the original watermarking is more than 96%,the algorithm has strong robustness.The parameter will be used to modify the singular values of a matrix according to some proportions and rules,which enhances the robustness of the algorithm,eliminates the false-positive and diagonal distortion problems of SVD applied to watermarking algorithm.