【作者】 胡金龙；

【导师】 刘九芬；

【作者基本信息】 解放军信息工程大学 ， 密码学， 2011， 硕士

【摘要】 随着网络技术、通信技术和多媒体信号处理技术的迅速发展及其应用的深入,信息安全问题日益突出。信息隐藏作为一种新兴的密码技术,已成为多媒体信息安全领域研究的一个热点。由于互联网上海量的文本数据、以及扫描成文本图像的个人档案和学历证明等,在人们日常生活中起到越来越重要的作用,文本数据的存储管理、产权保护及内容验证等安全问题成为信息安全领域的一个重要问题。针对文本信息隐藏技术的发展及其应用需求,本文主要研究文本信息隐藏算法的设计,研究结果如下:1.提出一种基于汉字部件的文本Hash算法针对汉字的特征,将汉字拆分为部件,提出一种基于汉字部件的文本Hash,并应用于文本零水印的构造和信息处理。实验结果表明:应用该算法构造的零水印具有较强的稳健性;而该算法应用于信息处理时具有较好的过滤效果。2.提出一种基于线性移位寄存器的文本水印算法首先利用LSFR序列的性质,构造一种适用于一维数据载体的水印信息模型;然后根据文本标点符号的编码特点及显示效果,提出一种信息嵌入方法;最后使用该方法将经水印信息模型产生的水印信息嵌入载体。该算法不增加文本数据占用的存储空间,不改变载体的文字内容,充分维护原创。实验结果表明:利用载体的任意一段都能恢复水印,且在较大的破坏力度下,仍能成功检测水印,具有较强的稳健性。3.提出一种基于块匹配和隐写码的灰度文本图像隐写方案(1)分析文本图像的特点,提出视觉同义块、视觉近义块以及文字内容语义稳健性的概念;提出一种多元码转换算法,该算法具有较高的转换效率;(2)利用同义块和近义块通过载体文本图像对秘密文本图像进行块匹配,将匹配块的位置信息作为嵌入信息;使用提出的转换算法将位置信息转换为多元码,进一步提高多元码的实际隐藏容量;为实现对任意嵌入率的编码,将几种隐写编码进行线性组合,利用隐写编码提高嵌入效率,增强算法的安全性。实验结果表明:在将文本图像作为隐藏对象时,该方案产生的秘密图像与原秘密图像具有较高的视觉相似性;同目前的大容量隐藏算法相比,该方案的安全性能和嵌入容量都得到较大的提高。更多还原

【Abstract】 With the rapid development and in-depth application of the network technology, communication technology and multimedia signal processing, the problem of information security is increasingly salient. As a new cryptographic technique, data hiding becomes a cutting-edge of the multimedia information security area. There is huge text data on the internet and text images scanned from personal document, educational certification and so on, which play a more and more important role in everyday life. Thus,security problems become a crucial issue in information security area, such as storage management, intellectual property protection and content validation.Targeting on the development and application requirements of the text information hiding technology, the bodies of the studies would be outlined as follows:1. Proposing a text hashing based on components of Chinese characters. We split Chinese characters into components by its structure, and propose a text hashing based on components of Chinese characters. It is robust against attacks when used in text zero-watermarking,such as copying,reversing,removing, and synonym substitution, etc, and it performs a strong function when used in information processing.2. Proposing a scheme of digital text watermarking based on LSFR. First, through analyzing text coding and the property of LSFR, a digital text watermarking model based on LSFR is proposed, which applied to single dimensional data. Second, we propose a new method used of embedding information based on space and punctuation. At last, we use the method to embed watermarking message which was produced by the watermarking model in the text. There’s no need to change character contents for protecting the authorship, and it doesn’t increase bytes of the text. The simulation results show that, any part of the text can recover the watermarking message and has good robustness against attacks.3. Proposing a text image hiding scheme based on block-matching and steganographic codes.(1) Through analyzing the characteristics of text image, we present the concept of semantic robustness, synonymous block and almost synonymous block in text image; and then a conversion algorithm between binary and polynary system is presented, the algorithm has higher conversion efficiency.(2) We use carrier text image to match synonymous block and almost synonymous block for secret text image, and embed the position information of matched block into carrier text image. Then we use the conversion algorithm to switch the position information into plurality codes, which make further improvement on the embedding capacity. We combine some steganographic codes to obtain steganographic codes in any embedding capacities, which could improve the embedding efficiency, i.e. improve the secure performance.The experimental results show that when the scheme is used for hiding text image, the extracted secret text image is similar to the original secret text image. Comparing with some methods with large embedding capacity, the scheme has great improvement on the secure performance and embedding capacity.更多还原