Video Watermarking Algorithm In H.264 Compressed Domain By Li Congying

Under the rapid development of the Internet,video has become one of the most popular information dissemination carriers.With the growing of the video data on internet,the improper use of the video is raising.Therefore,people's demand for video copyright protection and video authentication is increasing,and related research on video watermarking technology is particularly important.H.264 has become one of the most widely used video coding methods because of its greatly improved coding efficiency compared with the previous coding method.The research of H.264-based video watermarking algorithm has high application value.According to the operation scope of the watermarking algorithm,the watermarking algorithm can be divided into two categories: a pixel domain algorithm and a compression domain algorithm.The pixel domain algorithm using previous image watermarking algorithm to embed watermark in original video source directly,then compresses to generate the compressed code stream.The advantage of the pixel domain algorithm is that this kind of algorithm can get good result because of the mature of the image watermarking algorithm.The compressed domain algorithm combines the video watermarking algorithm with various video encoding methods,and performs the watermark embedding operation in the compression encoding process.The advantage of the compressed domain algorithm is that it combines the characteristics of the video encoding method,and can directly embed the watermark in the encoding process.Based on the research of H.264 video watermarking algorithm in recent years,for the weak performance on anti-recompression and anti-requantize,this paper design two H.264-based robust watermarking algorithm.The two algorithms are both compressed domain algorithm.The main work is as follow:1.A moving target detection algorithm based on compressed domain is introduced.The main thought of the algorithm is: by calculating the threshold of all the motion vectors in each macroblock to obtain the motion threshold of the current macroblock.Then compared with the ideal threshold obtained by the experiment to divide the macroblock into motion blocks and Non-moving block.The motion target is further accurately located by the macroblock state around the current macroblock.This algorithm serves as the basis of the following two watermarking algorithms and can improve the robustness of the watermarking algorithm.2.Aiming at the lack of anti-recompression capability of video watermarking,a robust watermarking algorithm based on codeword replacement is improved.The main thought of the algorithm is: by appropriately replacing some parts of the codewords in the entropy coding to perform the watermark embedding operation.Using the moving target detection algorithm to detect the motion macroblocks before embedding the watermark.Experiments show that the algorithm does not change the bit rate of the video,and has little impact on the quality of the video.It can also resist the recompression attack.3.A robust watermarking algorithm based on energy difference is proposed.The main idea of the algorithm is: embedding the watermark by the relationship of the energy value of the sub-macroblocks in each macroblock.The energy value of the submacroblock means the number of the bits that in each sub-macroblock.Using the moving target detection algorithm to detect the motion macroblocks before embedding the watermark.This algorithm is mainly proposed to solve the problem of video watermarking against requantize attack.Experiments show that the algorithm is robust to requantize attacks.