【作者】 葛龙；

【导师】 王执铨；

【作者基本信息】 南京理工大学 ， 系统工程， 2006， 硕士

【摘要】 随着通信技术、信号与信息处理技术的迅速发展，保护信息安全的关键——加密技术越来越受到人们的高度重视。 由于多媒体信息具有直观性、高效性等优点，业已成为通信领域的新宠。但是，在犹如政治、经济、军事等场合，敏感的多媒体信息通过公开信道传输或者存储于面向公众的数据库中，很容易受到信息窃取、数据篡改等攻击。因此，对多媒体信息的安全加密研究变得越来越重要。多媒体信息具有数据量大、对加／解密速度要求高、通常需要编码压缩后传输或存储等特点，传统的加／解密算法应用于多媒体信息加密受到很多的限制，因此，亟待研究新型的适合于多媒体信息加密的算法。 本文从信息安全的角度出发，以近年来非线性动力学中混沌和神经网络的研究成果为理论基础，研究了适合于多媒体信息的加／解密算法。 首先，概括总结了密码学的相关理论基础，特别分类研究了适用于多媒体数据加密的算法及其特点。 其次，在分析利用混沌系统生成序列密码研究现状的基础之上，提出了一种利用时空混沌生成序列密码的改进算法，并在验证所生成序列具有良好伪随机性优点之后，将其用于图像加密，该方法计算简单，比低维混沌系统构造的密码具有更高的隐蔽性能。 再次，为了改善Hash函数性能，提出了一种利用混沌神经网络构造单向Hash函数的算法，并以文本数据为例，详细讨论了利用该方法所构造Hash函数的初值敏感性、密钥敏感性、单向性以及弱碰撞性等优点。 最后，针对多媒体数据的特点，提出了一种基于参数敏感性检验的选择加密方案，并以音频为例，验证了该加密方案所具有的抗攻击能力、安全性、快速性等优点。更多还原

【Abstract】 With the development of communication, signal and information processing technology, encryption thchnology, the key to protect information security, has already attracted more and more attention.Multimedia has become great favorite in communication, as a result of its excellent characteristics of intuition and efficiency. But, in some applicaton fields, such as political, economical and military affairs, etc., sensitive multimedia is transmitted by public channel or storaged in public data base. Which makes it face threaten of wire tapping or tamper. So, multimedia information encryption becomes more and more important. Usually, multimedia is of generous data, needing high speed in encryption or decryption, going with coding and decoding during transmitted or storaged. This makes traditional encryption algorithms restricted, when applied to multimedia. Therefore, new encryption algorithms fit for multimedia are deadly desiderated.From the point of view of information security, this thesis investigates new algorithms for multimedia encryption and decryption, based on recent theoretical result of chaos and neural network in nonlinear dynamics.Firstly, cryptology and related theory is summarized. Especially, multimedia encryption algorithms are classified and their characteristics are studied in detail.Secondly, surveying on the state of the art, a modified method to create streaming cipher based on spatiotemporal chaos is proposed. After tested performance, this cipher is used to encrypt image data. The result shows that this method has a stronger secret performance than the streaming cipher based on low-dimensional chaotic maps, and it is simple to be realized.Thirdly, a new method to construct one-way HASH function based on CNN (Chaotic Nerual Netwok) is proposed. Experiment result obtained from text data shows that the HASH function based on CNN has advantages of sensitivity to initial values and user key, irreversibility and weak collision.In the end, a selective encryption scheme based on sensitive data detect is proposed, aiming at characteristics of multimedia. Then, experiment on audio data has been done, which result shows that this scheme has advantages of resisting attack, high security and speediness.更多还原