Based Contourlet Transform And Nonsubsampled Contourlet Transform Watermarking Scheme

With the popularization of computer and the rapid development of the Internet, people enjoy sharing world full of fun, through various search engines and network tools, we can very convenient to get the information. Facing such a good environment, some people will use this convenient to do some illegal behavior. For example common piracy, enterprise's important commercial secrets theft behavior, etc. When problems arise, then this way to tackle. In this context, the digital watermarking technique. The main work is to study how various network attack against behavior. Following is the study content:1. Based on the normalized and digital watermarking scheme Contourlet transform. The main idea of the algorithm is Contourlet transformation in the low finding the most appropriate location embedding watermarks, considering the image quality and robustness of the problems in the fold, so we also introduced the concept of human visual system. Finally, according to the human visual masking.watermark, through the parity of coefficient quantification modulation outline will be a size for 256-bits the image watermarking embedding readable. Due to the contour transform themselves have very good anti-noise ability and compression, i.e. normal attack ability strong resistance. But in order to algorithm is more perfect, more comprehensive ability against the attack, we also combines the normalization method. So both advantages complementary through. This algorithm of all-round attack ability strong resistance.2. Watermarking Scheme Based on Nonsubsampled Contourlet transform OF Pseudo-Zernike Moments. Nonsubsampled Contourlet transformation is a sample of new mathematical transformation, it is multi-purpose in image fusion or image denoising. It has good resistance to shift and antinoise sex. This advantage we could put it into the image watermarking. The main idea is to use the algorithm Nonsubsampled Contourlet transformation frequency conversion extract, because in the low frequency area compressed very well. So the ability to attack this algorithm has very comprehensive. But in order to achieve better effect, this paper also introduced against false Zernike torque this method. Take advantage of this algorithm combining two after the test result very ideal.