【作者】 周庆；

【导师】 廖晓峰；

【作者基本信息】 重庆大学 ， 计算机应用技术， 2008， 博士

【摘要】 随着计算机和网络技术的发展,多媒体已被广泛地运用到人们社会生活中的各个领域。同时,多媒体通信的安全问题也日益受到关注,而密码学是解决这一问题的首要技术。与文本数据不同,多媒体的数据量往往非常庞大,对其加密会消耗大量的计算资源和能量。数字图像是多媒体的一种,其研究成果能够较容易推广到其它多媒体形式中。本论文主要研究在保证安全要求的前提下,如何利用数字图像的特性实现快速加密。论文的工作主要包括五个方面:①提出了一种空域图像的并行加密模型。以往的空域图像加密模型中均采用密文分组链接的方式,这使得图像加密算法只能串行实现。为了提高图像加密的速度,论文提出了一种并行图像加密模型,并基于该模型给出了两个具体的图像加密算法。②提出了两种数字图像无损压缩技术即四叉树编码和MH编码的快速加密算法,其中重点研究了MH编码加密算法的性能。现有的多媒体加密技术在研究加密算法的性能时主要采用定性分析的方法,本论文在考查MH加密算法时对加密算法的压缩性能、视觉质量改变量和加密数据量进行了定量分析。③提出了两种最流行的数字图像有损压缩标准GIF和JPEG的快速加密算法。重点研究了对JPEG标准的加密技术,针对保密加密和退化加密两种不同的加密模式提出了四种加密算法,并对算法的各种加密性能进行了量化分析;提出了基于图像处理技术的新的安全性分析方法。④对一种基于多Huffman树的多媒体加密算法进行了密码学分析。基于多Huffman树的多媒体加密算法在图像加密领域受到了广泛的关注,本论文从选择明文攻击、已知明文攻击和唯密文攻击等多个角度对该算法进行了详细的密码学分析,指出该算法由于安全方面的缺陷不适合在实际中使用。⑤将图像快速加密技术的研究成果应用到MPEG流媒体的加密中。与图像相比,视频的快速加密技术具有更重要的意义,但两者又有密切的联系。本论文将图像快速加密技术的研究成果应用到MPEG的加密中,并研究了流媒体加密的困难及其解决方法。最后是全文的总结,并给出了论文的不足之处以及今后进一步研究的方向。更多还原

【Abstract】 With the development of computing and network technology, multimedia data are used more widely. At the same time, the information security for multimedia data becomes increasingly important and cryptography is the primary method to resolve it. Contrast to texts, multimedia contents contain usually huge volume of data, which make their encryption become very slow and power-consuming. This paper focuses on the approaches to improve the speed of encryption based on the characteristics of digital images, under the condition that those approaches satisfy the security requirements. Some of the conclusions of this paper are listed as follows.First, a parallel image encryption model in space domain is proposed. It is well known that cipher block chain mode is widely used in almost all image encryption algorithms, which makes the encryption can not be implemented effectively in parallel computing platform. To fast the encryption, we proposed a parallel image encryption. Two image encryption algorithms are proposed based on this model too.Secondly, two fast image encryption algorithms are proposed based on lossless compression techniques namely quad-tree coding and MH coding. So far, the performances of all known algorithms are checked by qualitative analysis and simulations. In this paper, the encryption algorithm for MH coding is designed so that its performance can be analyzed quantitatively by using information theories.Thirdly, two fast image encryption algorithms are proposed for two most popular loss compression standards GIF and JPEG. Four encryption algorithms are proposed for JPEG standard, covering both confidentiality and degradation mode for image encryption. The performances of JPEG encryption are analyzed quantitatively too. Furthermore, a new cryptanalysis method is proposed based on image processing techniques.Fourthly, a multimedia encryption algorithm based on multiple Huffman table is crypto-analyzed. This well-known algorithm is analyzed by chosen-plaintext attack, known-plaintext attack and ciphertext-only attack respectively in this paper. It has been pointed out that this algorithm is not suitable for practical use.Fifthly, the above results are applied to video encryption, which is of greater value than image encryption. The encryption algorithms for JPEG standard are revised and applied to the encryption of MPEG. The problems and their solutions when encrypting the stream media are also discussed in detail.Finally, the dissertation is concluded. Some problems as well as further work are also given.更多还原