Research On Image Watermarking Based On Lifting Scheme Wavelets

With the fast development of computer network, information model on the transmission has undergone enormous changes, the exchange of information has become more efficient and the form of information has been diversified. Especially, digital images are more and more prevalent. Because digital information can be copied and modified easily,a lot of piracy and copyright issues exist in the digital information that transmitted in the net work. Therefore the protection of intellectual property has become more and more important. Digital watermarking as a new information security has attracted people's attention.Wavelets have found wide applications in image processing due to its multiresolution and multiscale analysis properties. In addition, the aspect of wavelets is helpful in managing a good distribution or location of the message in the image in terms of robustness versus invisibility. Lifting wavelet is a new method in wavelet construction. Its characteristic is that all computations are processed in spacial domain, independent of frequency domain. Without relying on Fourier analysis, the transformation process can be performed with rapid algorithm, so it's widely used in graphics and video processing.This thesis focuses on lifting schemes of wavelets and their application algorithms for image digital watermarking. The main contents are outlined as follows:Firstly, we introduce the essential features, principle and the state of the art of the digital watermark, including essential feature and classification of the digital watermarking, sum up the basic frame of the digital watermark system, and introduce the common digital watermarking algorithms at present.Secondly, Human Visual System (HVS), fundamental principle of wavelets, Mallat algorithm and two-dimensional wavelet decomposition of image are analyzed.Then we introduce the algorithm of lifting wavelet and its applications in image digital watermarking, and propose a new watermarking method based on luminance identity. According to human virtual system, the lighter or darker the original image is, the bigger the contrast sensitivity threshold is. We can insert watermarks into the most lightful area of the image based on luminance identity. This method combines area space and frequency space effectively. And the validity of the algorithm is verified. with the artificial experiment.Afterwards, an adaptive color image watermarking algorithm is proposed, which is based on block classification and lifting scheme wavelet. First we change the original image from RGB color space to YUV color space, then extract Y component and split it into blocks. According to the visual masking features of HVS, watermarking insert scale factor of every block was calculated. Based on the lifting scheme wavelet, watermarking components were inserted into the largest coefficient of low-frequency coefficients adaptively according to the masked scales. The experimental results demonstrate that this method can be carried out simply and has good robusticity.