Research On Digital Image Watermarking Algorithm Based On Nonsampling Contourlet

Digital watermark technology is an important means for protecting themultimedia information, strike an average between transparency and robustness is animportant question for researchers. At present, the researchers propose manywatermarking algorithm in transform domain, such as the watearmarking basded onDCT，DWT and so on, and these algorithm split the difference between thetransparency and robustness. This paper focuses on the existing algorithms based ontransform domain, does some research on watermarking algorithm byNSCT(Nonsampling Contourlet Transform) tramsformation, singular valuedecomposition and jitter quantitative method。This article mainly as follows:1. Introduce the DCT and DWT transform’s structure and characters, propose anaudio watermarking algorithm based on DWT, and compare with the audiowatermarking algorithm based on DCT by simulation. By means of the analysis thealgorithms, could see that the audio watermarking algorithm based on DWT has morerobustness on resisting the attacks of shear, noise, filter and compression than DCT.2. Recommend the principle and characteristics of NSCT and SVD, analysis thepossibility and advantage of the algorithm, which is generated by the NSCT and SVD.Raise a watermarking algorithm based on SVD in NSCT and DCT’s dual transformdomain, by the NSCT’s sub-band of low-frequency has the same size with the orginalimage and the coefficient are some relationships.Transform the original image byNSCT, do the8*8block DCT transform with the low-frequency sub-band and thendecomposite the new rand with SVD, then the watermark is also on SVDdecomposition, overlay the two singular value decomposition matrix in order toembed the watermarking. The simulation has proved that the new algorithm is notonly good at in advancing the capacity of embedding watermarking, but also has goodrobustness in resistenting the Gaussian filtering, average filtering and the rotationattack. 3. Introduce the chaos scrambling transform and jitter quantification theory, putforward a bland watermarking algorithm based on NSCT and jitter quan-tification.First, transform the image with NSCT and block SVD decomposition, then scramblethe original watermarking with chaos in order to eliminate the linear correlation.Subsequently, change the every block’s singular value by jitter quanti-fication inorder to embed the watermarking. The algorithm is a bland watermarking algorithmas extract the watermarking doesn’t need the original image. Experimental’s resultsshow that this algorithm has a good transparency and strong robustness, and realize ablind extraction in watermarking.