基于视觉感知特性的频域数字水印研究

发布时间：2018-05-13 15:09

本文选题：信息隐藏 + 数字水印　；参考：《中北大学》2014年硕士论文

【摘要】：随着互联网技术的不断发展和普及，信息技术为多媒体信息的收发和存取提供了极大的便利，也使信息之间的交流达到了从未有过的广度和深度。信息隐藏利用其伪装这一特点在信息安全技术中应运而生，并成为隐蔽通信和知识产权保护等的有效手段，在国际上得到广泛的研究与应用。在信息安全保障体系中的应用主要集中在身份认证、数字作品的版权保护与盗版追踪、数据保密通信、数字作品的完整性、真实性鉴定与内容恢复等。数字水印，作为多媒体安全领域中隐藏技术不可或缺的一部分，是实现跟踪侵权行为和有效保障数字作品版权保护行之有效的方案，其固有的特点还有秘密通信、隐含标注和数字文件的真伪鉴别等，是目前学习和研发应用的一个重点。本文在分析已有的数字水印模型的基础上，以模型的建立和水印算法作为研究重点，针对水印嵌入容量、鲁棒性和不可见性等方面，结合HVS和灰度图像在频域中的一些特性，对基于图像的自适应水印嵌入进行了深入的研究，并构建了数字图像水印的系统模型，使得水印容量增加，，系统特性提高，抗攻击能力增强，并且实现了水印的盲提取。本文研究的主要内容如下： 1.针对空域中嵌入的水印鲁棒性差、易破坏和易破解等问题，本文基于频域提出了一种鲁棒性强、不可见性好的数字水印模型。分析预测水印系统的性能，实验仿真和结果表明了该模型的有效性和可行性，适用于图像水印的嵌入。 2.利用混沌映射或Zigzag扫描技术对原始水印信息进行置乱、维度变换等，再通过特定的公式计算来对水印进行加密保护，以此增强水印信息的安全性，即使图像被破坏者检测到含有水印信息，也不可能真正的被提取出来。 3.传统的数字水印模型没有从人眼视觉角度来考虑嵌入后图像质量的优劣，本文在深入分析数字图像结构和人类视觉模型的基本原理后，将二者相结合并获取阈值来达到水印嵌入的目的，最终提高了水印的不可见性。 4.频域变换后的图像系数按高低频带分布，本文通过对已有文献的参考比较，针对不同频域变换和内容3中的阈值自适应选择不同的频带和不同的系数位作为水印的嵌入位置，嵌入容量也有所增大。 5.本文利用奇偶量化的特性，对阈值进行特殊处理，提出了水印的嵌入算法，建立了最终的数字水印模型。本模型在水印的提取中不需要使用原始载体图和水印图，实现了水印的盲提取，同时提高了实验的运算速率。
[Abstract]:With the continuous development and popularization of Internet technology, information technology provides great convenience for the transmission and access of multimedia information, and also makes the communication between information reach unprecedented breadth and depth. Information hiding, which uses its camouflage, has emerged as the times require in information security technology, and has become an effective means of covert communication and intellectual property protection, and has been widely studied and applied in the world. The applications in the information security system are mainly focused on identity authentication, copyright protection and piracy tracing of digital works, data security communication, integrity of digital works, authenticity identification and content recovery, etc. Digital watermarking, as an indispensable part of the hiding technology in the field of multimedia security, is an effective scheme to track infringement and effectively protect copyright of digital works. Implicit tagging and digital document authentication is a key point in current research and development. Based on the analysis of the existing digital watermarking models, this paper focuses on the establishment of the model and the watermarking algorithm. Aiming at the embedding capacity, robustness and invisibility of the watermark, this paper combines some characteristics of HVS and gray image in the frequency domain. In this paper, the adaptive watermark embedding based on image is deeply studied, and the system model of digital image watermarking is constructed, which makes the watermark capacity increase, the system characteristic increases, the anti-attack ability is enhanced, and the blind extraction of watermark is realized. The main contents of this paper are as follows: 1. Aiming at the problems of poor robustness, easy to destroy and easy to crack embedded watermark in spatial domain, this paper proposes a robust and invisible digital watermarking model based on frequency domain. The performance of the predictive watermarking system is analyzed, and the experimental results show that the model is effective and feasible, and it is suitable for image watermarking embedding. 2. Chaotic mapping or Zigzag scanning technology is used to scramble the original watermark information, transform the dimension and so on, and then encrypt and protect the watermark through the calculation of specific formulas, so as to enhance the security of the watermark information. Even if the image is detected with watermark, it is impossible to extract it. 3. The traditional digital watermarking model does not consider the quality of embedded image from the point of view of human vision. In this paper, the structure of digital image and the basic principles of human visual model are deeply analyzed. Finally, the invisibility of the watermark is improved by combining the two and obtaining the threshold value to achieve the purpose of watermark embedding. 4. The image coefficients of the frequency domain transform are distributed according to the high and low frequency band. This paper adaptively selects different frequency bands and different coefficients bits as the embedding position of the watermark according to the different frequency domain transform and the threshold value of content 3. The embedding capacity is also increased. 5. In this paper, we use the characteristic of parity quantization to deal with the threshold value, put forward the embedding algorithm of watermark, and establish the final digital watermarking model. In this model, the original carrier image and watermark image are not used in the watermark extraction, so the blind extraction of watermark is realized, and the experimental operation rate is improved at the same time.
【学位授予单位】：中北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP309.7

【参考文献】