Digital Video Watermarking Algorithm Based On Transform Domain Research

With the rapid development of information technology and internet technology, various forms of multimedia works (image, audio, video, etc.) can be transfered through a variety of communication networks conveniently. However, digital media is also facing with severe security problems. Multimedia data can be edited, modified and copied arbitrarily in the transmission process, so copyrights of these works are being challenged. At the same time, the credibility of information transmitted by digital multimedia has also become a new problem. Digital video watermarking technology solved this problem well, it can protect copyright and authenticate content of digital works, and thus it has become a hot topic in the current academic research field.Analysis based on the discrete cosine transform and discrete wavelet transform, combined with the human visual system (HVS), two transform domain-based digital video watermarking algorithms were designed in this paper. The main work of this paper was as follows:(1) Introduced application background, current research status, basic models and algorithms classification of video watermarking technology, focused on the existing digital video watermarking algorithms, and these above knowledge could provide a theoretical basis of the design work of algorithm in this paper.(2) An adaptive blind video watermarking algorithm based on discrete cosine transform was designed. Firstly, analyzed the difference between the DC coefficient and AC coefficient of the discrete cosine transform, considering of the balance between imperceptibility and robustness, chose intermediate frequency region of the exchange coefficient to embed watermarking. And then,8×8DCT was taken on the luminance component of the video frame, and divided small blocks into the complex texture areas and the smooth texture areas depending on their variance. Finally, based on theoretical knowledge of the human visual system, chose a larger embedding strength in the complex texture areas, and chose a relatively small embedding strength in the smooth area, the watermarking information was embedded adaptively.(3) An adaptive blind video watermarking algorithm based on discrete wavelet transform was designed. Firstly, in order to improve the security of watermarking information, the embedded watermarking image was scrambled by Arnold transform, and then three layer discrete wavelet decomposition was taken on the luminance component of each frame of the video sequence, the embedding strength was chosen adaptively based on the energy of the wavelet coefficients carrying in the low frequency region, and then embedded encrypted watermarking information into the low frequency coefficients of wavelet transform. Watermarking extraction doesn’t need original video signal and original watermarking, realized blind extraction. (4) The two video watermarking algorithms designed in the paper were simulated using Matlab, and the experimental results show that the two digital video watermarking algorithms both have high security and good imperceptibility, and they can against attacks such as noise interference, frame deletion, frame shear, filtering, and so on.