AVS视频水印技术的研究和实现-智能与分布计算实验室

AVS视频水印技术的研究和实现

姓名	成楼
论文答辩日期	2007.05.31
论文提交日期	2007.06.07
论文级别	硕士
中文题名	AVS视频水印技术的研究和实现
英文题名	The Research and Implementation of Video Watermarking based on AVS standard
导师1	卢正鼎
导师2
中文关键词	视频水印;音视频编解码技术标准;差分数量水印算法;整数变换;指数哥伦布码
英文关键词	Video Watermarking;Audio Video Coding Standard (AVS);Differential Number Watermarking (DNW);Integer Transform;Exp-Golomb
中文文摘	随着计算机网络和多媒体(数字图像、视频、音频)技术的广泛应用，多媒体知识版权保护成为一个迫切需要解决的问题。数字水印技术作为解决这类问题的一种潜在的有效技术，正受到国内外学者的广泛关注。视频水印技术作为视频多媒体版权保护的重要手段得到了越来越广泛的研究与应用。首先，在研究和分析当前视频水印研究现状的基础上，结合视频水印的特点，视频水印的分类和当前主要的视频水印嵌入方案，深入讲解了现有的经典视频水印算法，指出了优缺点，论述了视频水印技术与视频压缩标准相结合的必然性。接着，全面分析了我国具有自主知识产权的国家视频标准音视频编解码技术标准（AVS），阐述了AVS的系统结构、解码流程和核心技术。接着跟MPEG系列视频标准进行对比，讲解了AVS标准的先进性，说明了基于AVS视频标准的水印研究的紧迫性和广阔的应用前景。重点分析了AVS标准在整数DCT变换，帧内预测和熵编码方面跟MPEG-2的不同，并结合AVS标准改进了DNW算法，设计和开发出了原型系统。在该改进算法中，将水印信息经置乱后嵌入到视频的亮度系数中，水印嵌入和提取时不需要完全解码，大大减少了运算的复杂度，提高了实时性。水印检测时不需要原始视频，实现了盲检测。另外由于嵌入到亮度类DCT的交流系数中，所以可在保证视频效果不失真的前提下，使水印具有很强的健壮性。由于跟AVS视频标准紧密结合，进一步提高了算法的鲁棒性。试验结果表明，水印嵌入前后，视频质量没有明显的变化，在无攻击时检测到的水印信息和原水印信息相比也没有明显的变化。方案中的水印能对抗帧跳跃和帧删除攻击，能应对一般的图像处理如剪切、滤波、锐化和加噪声等，都具有较好的鲁棒性。
英文文摘	With the rapid development of computer network and multimedia (digital images, video, audio) technology, multimedia knowledge of copyright protection has become an urgent need to address the problem. Digital watermarking technology as a solution to the problem of a potentially effective technology, by scholars at home and abroad is widespread concern. Video Watermarking technology as video multimedia copyright protection is an important means of increasing the extensive research and application. First, the research and analysis of the current status of video watermarking study on the basis of combining the features of video watermarking, Video Watermarking classification and current major video embedding program, in-depth briefing on the existing classic video watermarking algorithm pointed out the advantages and disadvantages discussed Video Watermarking technology and video compression standard combination of inevitability. Then comprehensive analysis of China's own intellectual property rights with the National Video Standard Audio Video Coding Standard (AVS), described the structure of the system, decoding process and the core technology. Then with a series of MPEG video standards contrast, the AVS on the advanced nature Note based on the standard AVS Video Watermarking study the urgency and broad application prospects. Analysis focused on the AVS standard integer DCT, intra prediction and entropy coding connection with the MPEG-2, AVS combined with improved DNW algorithm design and development of a practical system. The improved algorithm, watermark information by scrambling after I embedded video frame brightness coefficient, embedding and extraction need not completely decoded, greatly reduce the computational complexity, improved real-time. Watermark does not need the original video, realized the blind. As the brightness of the embedded category exchange DCT coefficients, and so we can ensure video effects without really premise, so that the watermark is very robust. As with the standard AVS Video closely integrated to further improve the robustness of the algorithm. Test results show that embedding around video quality has not changed significantly, in the absence of attack detection of the watermark and the original watermark than no obvious changes. Watermark program can jump confrontation frame and frame deletion of attacks to respond to the general image processing such as shearing, filtering, Sharpening processing noise has a better robustness.