The Study Of Audio And Image Watermarking Technology Based On Wavelet Transform

The rapid development of network technology provides much convenience to the spread of digital information. But, this convenience causes new problems. People can easily get audio and image files from the Internet. However, people can't make sure the integrity and authenticity of these digital works. Many tools can easily and successfully tamper audio and image files at present. So, copyright protection and image authentication are very important.Although there are many algorithms for copyright protection and image authentication, some problems about performance and security still need study further. Wavelet transform, which has a multi-resolution analytic character, is a time-scale (time-frequency) analytic method of information. Moreover, as both of its time-window and frequency-window can be changed it processes the ability of expressing local characters of information in time-frequency domain. A new audio watermarking algorithm, a new image fragile watermarking algorithm and a new semi-fragile watermarking algorithm in wavelet domain are proposed in this thesis. In audio robust watermark algorithm, watermark is embedded in transform domain by using the speediness, simplicity and multi-resolution of wavelet transform. When extracting watermarking, the audio is decomposed by Single-level discrete 1-D wavelet transform, and watermarking is extracted by ICA based on kurtosis maximization. This method's computational cost is low. Two image watermarking methods embed watermarks into transform domain using the multi-scale characteristic of wavelet. In fragile image watermarking algorithm, edge characteristics of image are embedded into original image by exploiting the characteristic of QR decomposition. This algorithm is sensitive to all kinds of attacks. In semi-fragile image watermarking algorithm, a robust watermark and a fragile watermark are extracted from the third approximation coefficient of wavelet transform of the original image. The two watermarks are embedded into the third vertical and horizontal detail coefficients. When authenticating image, image can be authenticated correctly and attacked parts can be localized through comparing the original watermarks and extracted watermarks.A lot of experimental results show that audio watermarking is robust. Furthermore, two image methods'exactness of tamper detection and location can...