基于H.264的数字视频水印算法研究

发布时间：2018-02-25 04:08

本文关键词： 数字视频水印 H.264 视频认证内容特征篡改定位　出处：《南京航空航天大学》2008年硕士论文　论文类型：学位论文

【摘要】： 数字视频水印是对视频内容进行版权保护和安全认证的有力工具,可以通过把信息嵌入到视频载体中去,实现版权保护、完整性认证、广告监控等功能。H.264作为新一代视频编解码标准,具有压缩性能高、应用范围广的优点,特别适合于低比特率下高质量视频服务。因此,基于H.264标准的视频水印认证方案的研究具有重要意义。本文对当前基于H.264编码标准的视频水印方案进行了研究,分析其存在的缺陷。然后从视频内容认证的特点出发,结合H.264的视频编解码系统,对基于能量差的视频特征提取和嵌入的方法进行改进,设计了一种基于H.264的内容认证方案,该方案在达到对视频内容认证要求的同时,实现了对篡改区域的检测和定位。通过对H.264帧内编码的DCT系数特性的深入研究,设计一种新的适用于低比特率视频基于压缩域的认证水印方案。在H.264视频编码时,根据I帧DCT低频量化系数关系生成特征码,以水印形式嵌入到I帧的DCT高频量化系数中。初步的实验表明,两种算法均有较小视频失真,计算复杂度低,较好的满足了视频水印的不可感知性、盲检测和实时性等要求,具有有较高的应用价值。
[Abstract]:Digital video watermarking is a powerful tool for copyright protection and security authentication of video content. As a new generation of video coding and decoding standard, advertising monitoring and other functions. H.264 has the advantages of high compression performance and wide application range, especially suitable for high quality video service at low bit rate. The research of video watermarking authentication scheme based on H.264 standard is of great significance. In this paper, the current video watermarking scheme based on H.264 coding standard is studied, and its defects are analyzed. Then, based on the characteristics of video content authentication, a video codec system based on H.264 is proposed. This paper improves the method of video feature extraction and embedding based on energy difference, and designs a content authentication scheme based on H.264, which not only meets the requirements of video content authentication, but also realizes the detection and localization of tampered regions. A new authentication watermarking scheme based on compressed domain for low bit rate video is designed by studying the characteristics of DCT coefficients in H.264 intra coding. In H.264 video coding, the signature is generated according to the DCT low frequency quantization coefficient relation of frame I. The watermark is embedded into the DCT high frequency quantization coefficient of I frame. The preliminary experiments show that the two algorithms have small video distortion and low computational complexity, which can meet the requirements of imperceptibility, blind detection and real-time of video watermarking, so they have high application value.
【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2008
【分类号】：TP309.7

【相似文献】