【作者】 孙德辉；

【导师】 单承赣；

【作者基本信息】 合肥工业大学 ， 信号与信息处理， 2007， 硕士

【摘要】 近年来随着数字视频在社会生活中的普及，视频作品遭到恶意攻击和非法侵权的问题也越来越严重。如何有效地对数字视频进行版权保护，成为多媒体技术发展中迫切而现实的问题。视频水印技术正是为解决这个问题而受到人们越来越多的关注，成为当前学术领域研究的热点。水印技术存在的基础是其技术的先进性，其归根结底是利用冗余来携带水印信息，而视频压缩的目的是去除冗余，因此好的水印设计方案应该充分考虑视频压缩编码的新动向，紧密结合当前及未来的标准进行研究。而H.264作为最新一代的视频编码标准，迫切要求水印技术与其相融合，提升和改进现有的视频水印算法，以适应新的应用环境。本文在对视频水印技术及H.264标准分析研究的基础上，提出了一种针对H.264低比特率码流特性的鲁棒视频水印设计算法。该算法首先将二值水印图像扫描为序列，并扩频处理，然后在H.264帧内预测编码阶段根据一定的判定条件，将水印序列逐比特地嵌入到Ⅰ帧每个宏块的16个4×4子块所选定的某个中、高频变换系数中，水印嵌入具有一定的自适应性，同时引入基于Lagrangian优化算法的编码控制模型，以提高含水印视频码流的率失真性能，最后对算法的嵌入和提取过程进行了仿真实现。实验结果表明，本算法在得到视频比特流较好率失真性能的同时对重量化编码、滤波、添加噪声和帧编辑等普通攻击具有一定的鲁棒性；水印提取时不需要对压缩码流完全解码，并且为盲提取；算法复杂度低，能够满足实时随机检测的需要，具有一定的实用性。更多还原

【Abstract】 In recent years, with popularity of digital video in social life, the issue of video works by malicious attacks and illegal use has become increasingly serious. How to protect the copyright of digital video effectively has become an urgent and realistic issue in the development of multimedia technique. Video watermarking technique to solve this problem is paid more and more attention by researchers and has emerged to a flourishing research topic in current academic domain.The basis for the existence of watermarking technique is the advanced nature of its technology, which uses redundant information to carry watermarking. And video compression aims to remove redundant, so a good watermarking scheme should take full account of the new video compression coding trends in close connection with the current and future standards for research. H.264, as the latest generation of video coding standard, has an urgent requirement of integration watermarking technique with it to enhance and improve the existing video watermarking algorithms to adapt to the new application environment.This paper proposes a robust video watermarking scheme aiming at H.264 low bit rate stream based on the analysis and research of video watermarking technique and H.264 standard. This scheme scans a binary watermark image into sequences firstly and processes them with spread-spectrum, and then embeds watermark sequences into certain median-frequency or high-frequency transform coefficients selected bit-by-bit by judgment under certain conditions in all 4×4 sub-blocks of each macroblock of I frames in H.264 intra prediction coding stage. The embedding process has certain adaptability, and introduces the coding control model based on the algorithm of Lagrangian optimization to enhance the rate-distortion performance of the watermarking video bitstream. Finally, experiments are performed to simulate the embedding and extraction scheme.Experimental results show that this algorithm obtains better rate-distortion performance of video bitstream, and has certain robustness against some common attacks such as requantization coding, filtering, adding noise, frame editing and so on. Watermark extraction does not need to decode compressed bitstream completely, which realizes the blind extraction; this algorithm which has a low complexity can meet the requirement of real-time random detecting, and has certain practicality.更多还原