Study On Video Watermarking In Transform Domain

With the increasing maturity of digital video technology, people can be very convenient through the digital video to communicate. But consequent is that some criminals use improper ways to illegally piracy, copy and tamper the video, which seriously infringe upon the rights of the copyright owner. Digital video watermarking technology has been proposed, can well ensure the video copyright issues, as well as to the digital video content authentication. Many people have paid more and more attention on the digital video watermarking technology which has become one of the hotspot of current multimedia information security technology.The frequency transform of digital video has various features, and can be a very good combined with the current popular video decoding standard. Therefore, the research on video watermarking in transform domain has a significant value. In this paper, the main research work is as follows:(1) Many previous video watermarking schemes embed the watermark by modifying the carrier video, which is very difficult to have a good trade-off between the invisibility and robustness. In this paper, a video zero-watermarking algorithm based on chaotic map and pseudo-3D discrete cosine transform (DCT) is designed. Firstly, the pseudo-3D DCT is performed on the original video frames, which are selected to construct the zero-watermark. Then, combining with chaotic map, zero-watermarking is generated by the sign sequence of pseudo-3D DCT coefficients. Finally, the zero-watermarking is uploaded to third party database. There is no modification for the original video, so, the proposed scheme has good invisibility. The experimental results show that it is robust against several attacks, such as H.264compression and MPEG-4compression, noisy attack, filtering, and frame shear.(2) Most of content authentication watermarking algorithms embed the watermark into the host video. The content authentication can be completed by comparing the original watermark and the extracted one. But this way need to transfer the original watermark, so the process of transmission is vulnerable, which increases the insecurity of watermarking algorithm. In this paper, a video watermarking authentication algorithm is designed based on the perceptual hashing technique. The perceptual hash value that represents the video content information is generated by the video transform domain coefficients of DWT and SVD, and then the encrypted perceptual hash value is as the watermark to be embedded into the low frequency DWT coefficient by the mean adjustment method. Thus, the perceptual hash sequence is included in the video which can represent the video’s content characteristics. In the watermark extracting process, the perceptual hash sequence is extracted to authenticate the integrity of the original video. The experimental results show that this algorithm has strong robustness, security, invisibility and certain ability of tamper localization.