Research On Digital Image Watermarking Algorithm Based On Multiwavelets Transform And Contourlet Transform

As the rapid development of Multimedia and network technology, the network has brought to people's lives convenient, but also has security implications the problem. Now, some people do dissemination, copying, tampering and other illegal operations for more convenience. Therefore, the multimedia information security has become a critical and challenging issue, which has been paid widespread attention at home and abroad. Therefore, digital watermarking as an important branch of information hiding, as a solution to multimedia security, authenticity and copyright security program is widespread concern and has become a rising international academic frontier.At present, some commonly used digital watermarking algorithms have a very good resistance to common noise, lossy compression, but the resistance to the combined attack of the anti-signal processing and geometric transformations is poor. Therefore, this article focuses on how to choose an appropriate transformation operation and embed watermark, which used to improve the robustness, resistance to combined attack of signal processing and geometric transformations. In this paper, using static digital image as the object, two digital watermarking algorithms based on transform domain have been designed.Multi-wavelet transform is also a wavelet transform, which has the features of symmetry, orthogonality, smoothness and compact support. In Digital Watermarking, Wavelet Transform has a strong filtering and compression anti-attack ability, but the wavelet transform does not have a rotation, scaling and translation and other geometric properties, so the geometric anti-attack ability of wavelet domain image watermarking is poor. Considering this point, this chapter extract the image features to resist geometric attacks. First of all, a scrambling encryption will be made in digital watermark image to increase the safety and robustness, then embedded watermark is brought to the the sub-scrambling encryption after a multi-wavelet transform. Finally it is the extraction of feature vectors .Experiments proves that the algorithm is not only highly robust in the noise and compression of conventional image processing, but also a very good resistance to geometric transformations.For the limitation of the wavelet transform in two-dimensional image, a block Energy Analysis Contourlet domain digital watermarking algorithm is proposed. Contourlet transform the most important is to get lines and surfaces with exotic images of sparse representation, it is not only inherited the wavelet multi-resolution time-frequency analysis feature, and has a good anisotropy. This algorithm is mainly from Contourlet factor analysis of the energy proceeding It firstly decompose the original image by Contourlet, and choose the low-frequency sub-band decomposed image as a watermark embedding sub-band, and then makes the low-frequency sub-band for non-overlapping sub-blocks, at the same time calculates each sub-block coefficients respectively. In selecting a large block of energy to embed the encrypted watermark scrambling, and then through the inversing transform of Contourlet , the watermarked image is received. These results prove that: in this chapter, the digital watermark embedding algorithm has a good visual effects, And it also has a good robustness of JPEG compression, scaling, smoothing filtering and other image processings.