Research On Video Streaming Watermarking

In recent years, the digital watermarking technology is a cutting-edge research in the international academic community.It is an important branch of information security and is closely linked with information hiding etc. By embedding secret information in the digital Products (Such as images, audio, video ) to prove copyright infringement or tracking. Because the digital product content is able to maintain a large number of replicated high fidelity, the pirated become very easy, So that the intellectual property is serious violations, the video stream processing of video data as the method of copyright protection is being taken seriously. Therefore,research on video streaming watermarking is very valuable both in academic development and potential applications. The concepts in this dissertation keep abreast of the international video streaming watermarking reserch progress. The main contribution of this dissertation is summarized as follows:(1) According to the new features of H.264/AVC coding standard,a video watermark algorithm based on AC coefficient is introduced.Firstly,to carrying out two-dimensional Arnold image scrambling to obtain the Irregular embedded images. Even if the attackers obtain the watermarking, it is difficult to obtain the original watermark. Robustness of the algorithm is improved.Then,embedded the watermark into AC coefficient quantized according to analysis the video frame pixel luminance and uniformity of texture. The watermark has good invisibility. The detection does not require original video sequence, so this algorithm is blind detection. The experimental results show that this watermark scheme can effectively resist the rotation and other basic geometric attack .Moreover it can resist lossy compression and noise, it gets better security and robustness.(2) Aiming at the problems of the features of 3G video coding standard and real-time, a robustness video watermark algorithm based on DC Coefficients of chroma block is presented. The two-dimensional Arnold image scrambling is designed firstly, according to calculating the brightness of the I frame video sequence to determine the embedded frame, then calculating the average value and variance of chroma block, embedded the scrambled watermark information into the quantized DC coefficients of chroma block so this algorithm is blind detection.Consider the intensity factors according to the human visual system, the watermark has good invisibility. The pretreatment of the watermark and the choice of embedded location enhanced the robustness in some extent. In the process of watermark embedding, the video do not need to decode and re-encoding, so it meets the requirements of real-time.(3) The emergence of video streaming expand the applications of the video streaming watermarking technology, but the applications of watermarking technology in streaming media are influenced by the bandwidth. In connection with the video streaming to apply to network transmission, this paper Proposed a watermarking technology, the watermark are embedded into the base layer and enhancement layer bitstream. In the case of low bandwidth, We can extract the watermark information from the basic layer to confirm the integrity of the video stream. In the case of sufficient bandwidth, we can extract the watermark from the enhancement layer. This paper use bit-plane DCT coding techniques, and then using linear scanning method encoded from the highest bit-plane block plane to the least significant bit plane. In the encoding process, select the target embedded watermark bit plane block, then the watermark is embedd into the 8×8plane block by modifying the specific coefficient. This algorithm not only has good invisibility and robustness as well as easy realized which meet real-time, but also can reduce the size of bit streams after embedding watermark information.