【作者】 周雪赞；

【导师】 童勤业； 张宏；

【作者基本信息】 浙江大学 ， 生物医学工程， 2004， 硕士

【摘要】 伴随着网络技术和多媒体技术的飞速发展，多媒体数据逐渐成为人们获取信息的重要来源，因此，如何保护多媒体信息的安全成为国际上研究的热门课题。数字水印技术作为信息隐藏技术研究领域的重要分支，也是版权保护的重要手段，已经受到越来越多的关注。它通过在原始数据中嵌入秘密信息——水印(watermark)来证实该数据的所有权或完整性，以此来抵制对数字作品的盗版或篡改等，在理想情况下，要求嵌入水印前后的信号不能有感觉上的差异，同时如果不损坏原始信号，水印能够准确提取或检测出来。本论文中，我们将研究主要集中在具有鲁棒性的不可见的有意义数字水印技术。分三部分工作进行： RS编码纠错原理在数字水印中的应用。利用RS码很强的纠错能力和较高的译码可靠性，对水印信息先进行RS编码，然后将编码后的信息(以图像形式表现)，分多次嵌入到水印载体图像中。根据嵌入方法，由其逆过程完成水印的提取工作，将提取后的水印图像根据纠错能力由RS译码恢复原始的水印信息。在这项工作中充分利用了RS强纠错能力，使得纠错能力最高可达47.66％，又通过选择纠错等级进一步增强了数字水印的鲁棒性，满足通信中数字水印的应用。 基于混沌序列的图像置乱算法处理水印图像。混沌映射有这样一些特征：(1)对初始条件敏感、因而是不可预测的；(2)非周期性。利用这些特征来置乱图案水印以保证水印的安全性。本文从空域和频域两个方面来置乱水印图像，再结合空域和DCT频域两种方法来嵌入水印，实验证明该数字水印技术有一定的抗压缩和抗随机噪声能力，将信息加密和信息隐藏很好地结合起来，满足实际需要。 用非线性函数迭代生成水印信息的数字水印技术。利用非线性函数迭代系统选择不同的初值和参数，迭代产生不同的序列形式，以此作为水印信息，完成有用信息的加密工作，也就是说，水印信息实际是一串非线性序列在图像中的表现，当序列受到局部损坏时，我们仍可把有用信息恢复出来，强调了水印的自恢复能力，进而根据人眼视觉特性选择载体图像的区域来嵌入水印，实验证实了该算法的有效性。更多还原

【Abstract】 The digital media has become a main way for information communication along with the wide use of network. At the same time, multimedia information security problems emerged as side effects of the popularity of digital representation and distribution over network. So the protection and enforcement of intellectual property for digital media has become an important issue. Watermarking which is one of the important fields of information hiding and the effective means to protect copyrights piracy is receiving increasing attention. It is the direct embedding of additional information into multimedia signal. Ideally, there should be no perceptible difference between the watermarked and original signal, and the watermark should be difficult to remove or alter without damaging the host signal. In this paper, we limit the scope of our research to robust and imperceptible digital watermarking. The main contributions of this dissertation are:In order to improve watermarking robustness against distortions and attacks, it is natural to envision the use of error correcting codes. According to the principle of RS, we encrypt the watermark which is an image. By limiting a different level to error permit, there is more difficulty in gaining the watermark when the destroyer doesn’t know the level and the data quantity. As for the embedded ways, we introduce the common way of embedding the watermark to the special domain of the image.A watermarked image encryption bases on chaos sequence. Chaotic sequences have several good properties including the ease of their generation, the sensitive dependence on their initial condition and the statistics property of noise. We implement the image encryption algorithm based on the technique of image scrambling transformation in special and frequency domain. And more, we embed the watermark to the special and frequency domain of the digital product. Preliminary results are satisfactory and efficient.Watermark is some information produced by the nonlinear function. That is to say, we encrypt the important information by repeating to use the nonlinear function.As a result, the watermark may be a sequence. The important information may be abstracted out of the watermark only in several sequence data. At the same time, we deal with the product to some block according to the human vision system and embed the watermark to blocks.更多还原