Research On Theories And Applications Of Wavelet And Digital Watermarking

This dissertation studies theories of wavelet and digital watermarking, investigates main problems with digital media, studies image lossless compression algorithms and image demosaicing algorithms based on wavelet theories through combination with those problems, and studies copyright protection methods for digital media based on digital watermarking theories. It studies the copyright protection method for genuine media from Genuine-media identification aspect and traitor tracing aspect.Main work and contributions of this dissertation are as follows:1. Theoretical research on wavelet transformFocuses on implementary theories of wavelet transform, which including implementary theories of real wavelet transform, complex wavelet transform, and analytic complex transform. The dissertation puts forward the design theory of all layers filters to implement analytic wavelet transform in dual-tree structure as soon as possible. It proposed two algorithms of implementing analytic wavelet transform, which named Q-shift FRA algorithm and common-factor FRA algorithm.2. Application research of IWT in the field of image lossless compressionThe compression performance of IWT with that of DWT is compared, moreover the advantages and disadvantages of IWT in practical applications is discussed. A novel lossless image compression scheme IWTAP is presented based on combining integer wavelet transform and adaptive prediction. The IWTAP improves the lossless image compression algorithm of JPEG2000 standard.3. Application research of lifting wavelet transform in the field of demosaicingFirst, a high-frequency similar model is presented on the integration of color difference model and wavelet transform theory, and then a condition is deduced based on theoretical analyses. In order to construct the model, this condition should be met in the process of wavelet transform. A method for reasonably integrating the characters of Bayer pattern with the process of wavelet transform is presented. If use this method to reconstruct R and B planes, it can greatly decrease the reconstructing time by avoiding iteration. The ratio of computational cost used by this method and used by the one which proposed by Gunturk, etc. is 19:432. Finally, a demosaicing algorithm is proposed. Compared with other typical algorithms, the above algorithm shows better on both demosaiciing performance and computational cost.4. Theoretical and application research of robust watermarking in the field of copyright protectionTheories of anti-collusion fingerprinting codes are researched systemically. Schemes of realizing digital fingerprinting in the process of copyright protection are studied base on combination with media transmission infrastructure, consumption patterns of user terminals, conditional access system, etc.5. Application research of fragile watermarking in the field of copyright protectionAn efficient genuine-video identification algorithm named as TDIA is proposed, which based on integration of fragile watermarking, video communication technology and video compression standards. It includes two parts: embedding algorithm and identification algorithm. Iterative method is introduced on the identification algorithms to eliminate the communications interference for the impact of watermarking. The performance of the iterative method is investigated according to the difference of nature between piracy operation and communications interference. Finally, the identification errors of TDIA algorithm are analyzed based on the results of iteration and by dint of the theory of Chernoff Bound. The total identification error of the algorithm is less than 10?1 8when BER (Bit Error Rate) of encoded video stream is not more than 10?3 and the total number of I-frames of Mpeg2 encoding video is more than 360.