【作者】 王琪；

【导师】 康宝生；

【作者基本信息】 西北大学 ， 计算机应用技术， 2016， 硕士

【摘要】 以数字水印作为各类数字产品版权保护的支撑技术正在迅猛发展,这使得针对水印技术的研究成为当前信息安全领域研究的热点。数字水印算法对各类常见攻击抵抗能力的强弱直接决定着该算法性能的优良与否。因此,增强算法的鲁棒性,提高算法的透明性,是水印研究领域的重中之重。本文在深入分析图像水印算法研究现状的基础上,综合运用MCA、RPCA和QR等理论的特性,对现有方法导致原始载体图像像素质量严重下降的问题和对常见攻击抵抗能力不强的问题展开了深入研究。本文研究工作的主要内容如下：(1)提出一种基于MCA的图像水印算法。针对水印算法易导致载体图像像素质量严重下降,且对抵抗常见攻击鲁棒性不强的问题,分层嵌入的思想被提出。利用MCA将图像分为平滑层和纹理层,可以减少对图像像素质量的影响；对纹理层进行DCT变换,选取变换后的中频部分嵌入水印,有助于提高算法的鲁棒性。数值结果表明,该算法不仅能获得较高的视觉质量,而且能十分有效地抵抗一定程度的裁切攻击,同时对椒盐噪声攻击和JPEG有损压缩攻击也都具有很强的抵抗能力。(2)提出一种基于RPCA的图像水印算法。针对(1)中算法对椒盐噪声攻击和部分裁切攻击抵抗能力不够强的问题,结合RPCA实现另一种分层嵌入的水印算法。对分层后的纹理部分进行DCT变换,并选取变换后的中频部分嵌入水印。最后合并所分层得到含水印图像。数值结果表明,该算法不仅能获得较高的视觉质量,且较(1)算法对抵抗椒盐噪声攻击和一定程度的裁切攻击具有更强的鲁棒性。(3)提出一种基于QR码的分块图像水印算法。针对大多数水印算法对大面积裁切攻击或磨损攻击抵抗能力不强,甚至不抵抗的问题,本文结合QR码的高容量和高纠错特性,将秘密信息编码成QR码作为水印。实验结果表明,该算法可以有效抵抗90%以上的裁切攻击,且对裁切攻击的抵抗能力与嵌入算法的执行次数成正比。同时该算法对抵抗高噪声方差的白噪声攻击、高斯噪声攻击和椒盐噪声攻击以及JPEG有损压缩攻击都具有很强的鲁棒性和较高的透明性。更多还原

【Abstract】 The technology that takes digital watermarking technology as the main copyright protection of digital products is growing rapidly, which makes the research on watermarking technology become a hotspot in the field of information security. Whether a watermarking algorithm is excellent or not is decided directly by its performance to resist various common attacks. Therefore, to enhance the robustness of algorithm and to improve the transparency of algorithm to all kinds of attacks is the priority among priorities in the field of digital watermarking research.In this thesis, the present situation of image watermarking algorithm is firstly intensive studied. Then aiming at the problems existing in the present situation of image watermarking algorithm, such as the existing methods will seriously decrease the pixel quality of the original carrier image and they have weak ability to resist the common attacks. They are carried out by synthetically using the characteristics of the theory of MCA, RPCA and QR.The main research work of this thesis is as follows:(1) A digital watermarking algorithm based on MCA is proposed. Aiming at the problem that the quality of the image pixels is reduced or damaged and the robustness of the watermarking algorithm is not strong, the idea of hierarchical embedding is put forward. The original carrier image is divided into smooth layer and texture layer by MCA, which can reduce the effect of image pixels quality. Take DCT transform on the texture layer and choose the intermediate frequency part to embed watermark, which can improve the robustness of algorithm. The results of experiment illustrate that it is not only can obtain higher visual quality but also improve the resistance ability of cutting attacks in a certain extent effectively. At the same time, it outperforms in robustness, such as salt and pepper noise attacks and JPEG compression attacks.(2) A digital watermarking algorithm based on robust principal component analysis is proposed. Aiming at the problem that the algorithm (1) has yet to be strengthened to salt and pepper noise attacks and cutting attacks in a certain extent, an alternative hierarchical embedded watermarking algorithm is proposed combinng with RPCA. Take DCT transform on the texture layer and embed watermark. Then, embedded the watermarking in the part of intermediate frequency and merge the two layers to get the watermark image. The results of experiment illustrate that this method not only can obtain higher visual quality but also has stronger robustness than (1) against the cut attacks in a certain extent and salt and pepper noise attack.(3) A digital watermarking algorithm, which can against large area cropping attacks or wear attacks is proposed. Aiming at the problem that the resistance ability of most of the algorithms for large area of cutting attack or wear is either not strong or even do not, this thesis combines the high capacity and high error correcting capability of QR code and propose a new algorithm:Digital watermarking algorithm for QR code based on DCT and blocked images. This algorithm encodes secret information into QR code as watermark. The results illustrate that this algorithm can effectively resist more than 90% of the cropping attacks to the maximum extent. And with the increase of embedding frequency, the resistance to cutting attacks is stronger. At the same time, it has strong robustness and high transparency to resist the white noise and Gauss noise, salt and pepper noise with high noise density and JPEG compression attacks with high compression factor.更多还原