Adaptive Equalization Watermarking Algorithm In Nonlinear Scale Space

The embedding location and intensity of the watermark seriously affect the performance of the watermarking algorithm.Analysis and comparison of various digital watermarking algorithms,the linear-scale spatial watermarking algorithm is not accurate enough to embed the watermark,and the embedding intensity parameter is random.As a result,the stability of the extracted feature points is poor,and the embedding strength parameters are random,and the workload of embedded watermark is increased.In order to solve the above problems,an adaptive watermarking algorithm based on nonlinear scale space is proposed to balance the transparency and robustness of the watermark.First,using KAZE algorithm to extract feature points in nonlinear scale space,and selecting stable feature points to construct embedded watermark region.Second,using three-level discrete wavelet transform to the region,and the diagonal matrix after the decomposition of the singular value of the low-frequency sub-band is used as the coefficient matrix of the watermark to be embedded.After the watermark image is scrambled,singular value decomposition is performed,and the product of right singular value matrix and diagonal matrix is used to remove false alarm.The new matrix is used as the embedded watermark carrier.Last,using the fitness function of FOA to calculate the embedding intensity,then superimposing the coefficient matrix on the watermark carrier,the embedding process is adaptive completed.The simulation results show that the peak signal-to-noise ratio(PSNR)of the watermark images is 44 dB,and most of them are 46 dB.After compression,noise,cutting and rotation attacks on the watermark carrier image,the normalized correlation coefficients between the extracted watermark image and the original watermark image are all up to 0.94,and some of them are close to 1.The results show that using the KAZE algorithm to select the embedded watermark information of the non-linear image block,locating embedding area is more accurate.Combining with the DWT and SVD algorithm is more stable.The FOA has high efficiency in adaptive selection of embedding watermark intensity parameters,which effectively balances the invisibility and robustness of the watermark algorithm.