基于压缩感知理论的稳健性视频水印技术研究

项目名称： 基于压缩感知理论的稳健性视频水印技术研究

项目编号： No.61272381

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 自动化技术、计算机技术

项目作者： 赵慧民

作者单位： 广东技术师范学院

项目金额： 80万元

中文摘要： 视频水印受压缩率、实时性和盲随机检测的限制，未解决的关键问题是在码率条件下隐藏数据的容量与稳健性的约束矛盾。图像水印算法最新引入的压缩感知理论，以嵌入数据的稀疏性和高质量信息的恢复性，受到信息隐藏领域的广泛关注。本申请将压缩感知应用于视频水印研究中，解决视频水印的数据量与实时性嵌入、盲随机检测问题,包括：（1）考虑视频图像的特征信息和稀疏表示的正交/非正交基，研究图像新的稀疏表示及其特征选择和提取方法，构造优化的冗余字典和测量矩阵,建立视频水印数据的生成模型。（2）在压缩域,根据压缩感知获得视频的不同特征，研究基于视频内容的多载波数据调制技术，设计各种攻击条件下的稳健性视频水印算法。（3）在空间域，研究压缩感知视频水印盲随机检测技术。（4）研究压缩感知视频水印实现的复杂性、稳健性和水印容量的关系，建立视频水印性能评价模型。项目成果将改善信息隐藏的质量,达到视频水印稳健性的研究目的。

中文关键词： 视频水印；压缩感知；稳健性；特征；

英文摘要： Due to limitation from compression rate and real-time requirement as well as random detection, key problems of the video watermarking are the confliction between data-hided capacity and robustness in rate control. At present, compressed sensing theory applied to information hiding has witnessed a growing interest, due to the aparsity of the embedding data and accuracy of the retrieval information. But the study was only used for image watermarking algorithms . This research project utilizes compressed sensing theory to robust video watermarking for solving some problems between real-time data embeding and blind random detection. It consists of four sections. Section I deals with image feature information and orthogonal basis of the novel data sparse representation, exploits image features section and extraction methods, establishes optimum redundant dictionaries and sensing matrix as well as novel video watermarking data models are established to information hiding. Section II, based on video various features at compressed domain, demonstrates the data embedded methods of multicarrier modulation technology by combining video content. Therefore, the technology would implement some robust video watermarking algorithms against various type of attacks. Section III, at spatial domain,investigates blind random dete

英文关键词： Video watermarking；Compressed Sensing；Robustness；Feature；