数字水印容量与能量研究

发布时间：2018-06-14 21:18

本文选题：数字水印 + 水印容量　；参考：《天津大学》2016年硕士论文

【摘要】：进入21世纪,信息化社会发展方兴未艾,信息安全问题日益突出。数字水印技术作为解决数字多媒体作品版权保护的一种得力手段,日益引起人们的关注。当前大多数的数字水印算法均是通过反复试验来确定嵌入数字水印的最佳容量和能量。论文针对这一局限性进行了以下主要研究:一、提出了填球水印模型,并推导出该水印模型下的容量公式以及容量与峰值信噪比之间的定量关系。该水印模型不需要对攻击信道作概率统计假设,而是将攻击行为量化为一个几何失真。二、研究了在一般意义下位替换水印模型容量的上下界公式,并对该水印模型下的容量进行了仿真。由于该容量公式为非解析形式,因此其结果可以用来估计容量大小。三、提出了向量内积水印,水印嵌入方式是基于叠加嵌入,水印提取方式是基于向量内积,并推导出该模型下的峰值信噪比公式。四、提出了改进的向量内积水印,并推导出该模型下的峰值信噪比公式、误码率公式。论文通过多组对比实验,证实了改进的向量内积水印对于抵抗高斯白噪声,幅值缩放、高斯滤波等常见攻击具有良好鲁棒性;引进数字图像置乱技术能有效抵抗剪切攻击对数字水印图像内容的破坏;引进特征点匹配能够有效抵抗几何攻击。本论文的研究,可应用于具体数字水印技术方案,具有一定的实用价值。
[Abstract]:In the 21 st century, the development of information society is in the ascendant, and the problem of information security is becoming more and more prominent. As a powerful method to solve copyright protection of digital multimedia works, digital watermarking technology has attracted more and more attention. At present, most digital watermarking algorithms determine the optimal capacity and energy of embedded digital watermarking by repeated experiments. The main research work is as follows: firstly, a watermark model is proposed, and the capacity formula and the quantitative relationship between capacity and peak signal-to-noise ratio are derived. The watermark model does not need to make probabilistic statistical assumptions for the attack channel, but quantifies the attack behavior into a geometric distortion. Secondly, the upper and lower bound formulas of the capacity of the watermarking model are studied in the general sense, and the capacity of the watermarking model is simulated. Since the capacity formula is a non-analytical form, its result can be used to estimate the size of capacity. Thirdly, a vector inner product watermarking is proposed, which is based on the superposition embedding, and the watermark extraction is based on the vector inner product. The formula of peak signal-to-noise ratio (PSNR) based on the model is derived. Fourthly, an improved vector inner product watermark is proposed, and the peak signal-to-noise ratio (PSNR) and error rate (BER) formulas are derived. In this paper, the improved vector inner product watermark is proved to be robust against Gao Si white noise, amplitude scaling, Gao Si filtering and other common attacks through a number of comparative experiments. The introduction of digital image scrambling technology can effectively resist the destruction of digital watermarking image content caused by clipping attack, and the feature point matching can effectively resist geometric attack. The research of this paper can be applied to the specific digital watermarking scheme, which has certain practical value.
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP309.7

【参考文献】