【作者】 师晶；

【导师】 林克正；

【作者基本信息】 哈尔滨理工大学 ， 计算机应用技术， 2008， 硕士

【摘要】 数字水印技术作为一门新兴的学科交叉的应用技术,是信息隐藏的重要分支。它是将标识作者版权的保护信息和认证信息嵌入到图像、音频、视频或软件等各种数字产品中,以达到区分非法拷贝传播和保护知识产权的作用。数字水印作为当前多媒体信息安全领域的热点技术,已经受到国际学术界和企业界的高度重视。本文首先结合国内外研究问题的背景,指出图像和视频水印技术对于科研及工业生产的重要意义,并且对水印技术的研究现状及发展趋势进行了分析和总结;然后综述了与数字水印和MPEG-4技术相关的一些基础理论知识,针对图像和视频水印技术进行了深入系统地研究。本文针对图像和MPEG-4视频分别给出了三种水印算法:(1)基于双变换的图像数字水印算法。将离散小波变换同时运用到载体和水印信号中,采取折中办法嵌入水印。该算法具有很好的半脆弱性;(2)基于DWT和SVD的改进分块图像水印算法。兼具离散小波变换和奇异值分解的双重性能,同时结合自适应块分解量化法分别进行灰度、彩色载体的水印嵌入。算法达到了高嵌入、高鲁棒性和高精确提取能力的统一结合;(3)基于视频帧的彩色水印嵌入算法研究。选择I帧当作彩色水印嵌入帧,此方案综合考虑了DCT、DWT及SVD的特性,根据纹理特性将水印信号自适应的嵌入到8×8分块的DCT系数中。此方案很容易被应用到MPEG-4视频码流中。实验结果表明,该算法具有良好的不可见性;同时对噪声、剪切等攻击都具有较强的鲁棒性和水印提取能力。本文的MPEG-4视频数字水印主要是针对未压缩域中彩色水印的嵌入技术,对视频产品的版权认证具有十分重要的意义。更多还原

【Abstract】 As being a new and interdisciplinary application technology, digital watermarking technology is an important branch of information hiding. It embeds the protected and authenticated information which has been labeled the author’s copyright into the kinds of digital products such as image, audio, video or software to distingiuish between unauthorized copy and transmission and protect intellectual property rights. Digital watermarking being a hot technology in present multimedia information security field has already been paid much attention by international academia and business circles.This thesis firstly pointed out the important meaning of image and video watermarking technology to the science research and industry production based on the background of problem studied, and analyzes the current status and developing trend of the digital watermarking. And then the basic principles, features and problems of watermarking and MPEG-4 were summarized. The image and video watermarking were researched thoroughly and deeply.Three algorithms were put forward in this thesis for image and MPEG-4 video: (1) Image digital watermarking algorithm was proposed based on double transformations. Put discrete wavelet transformation (DWT) into both carrier and watermark signal and embedded watermark with compromise. This algorithm has good semifrangibility. (2) An improved unblocking image watermarking algorithm based on DWT and singular values decomposition (SVD) was proposed. Utilized the properties of DWT and SVD simultaneously to embed watermark on both grey and color carriers with adaptive block decomposed quantification. This algorithm has reached high embedding quantity, well robustness and precise extraction ability. (3) Research on color watermarking algorithm based on video frames. I-frame was selected as the watermarking embedded frame. It combined discrete cosine transformation (DCT) and DWT’s character with SVD’s and embedded the watermarking signal into eight by eight blocks of DCT coefficients adaptively according to the texture property. This algorithm is easy to be used in MPEG-4 video. The experiment results indicate that it has well invisibility, strong robustness and extraction ability on nosing and cropping attacks.The MPEG-4 digital watermarking mentioned here is only for color watermarking technology in uncompressed video, and it has played a very impotant role on copyright authentication for video products.更多还原