基于同态加密的零知识水印安全检测方法研究

The widespread use of electronic documents provides a great convenience for people's daily work and life, on the other hand it also leads to many problems, such as copyright disputes, data leaks etc. Digital watermarking and data encryption technology play an important role to protect the security of such information.However, the security of the digital watermarking donot entirely depend on the algorithm or key itself. When detecting, the key of watermarking must be provided, once the key leaks, the attacker will attack the watermarking. Therefore, the security of watermarking is very important. This project will study the fully homomorphic encryption scheme which is simple and has smaller complexity, and which will be used in zero-knowledge watermark detection technology. The main reseach issues include: the fully homomorphic encryption scheme based on integers mode operation and algebraic number fields to improve the efficiency of the encryption schemes and enhance the practicability; public verifiable zero-knowledge watermark detection scheme based on multiplicative homomorphic encryption to prevent deception from the third party; non-interactive zero-knowledge watermark detection scheme based on the fully homomorphic encryption to improve the security of watermark detection. Therefore, the research not only establishes general model for zero-knowledge watermark detection, but also provides new technology and new method for the application of encryption and watermarking technology.

电子文档的广泛使用为人们的日常工作与生活提供了极大便利，同时也导致了如版权纠纷、数据泄密等问题。数字水印和数据加密技术为保障这些信息的安全起到了重要作用。然而数字水印的安全性不能完全依赖于水印算法或密钥本身。目前的水印技术在检测时必须出示密钥，而密钥一旦泄露，攻击者就会利用这些信息来攻击水印。因此水印安全性显得非常重要。本课题研究复杂度小且易实现的完全同态加密方案，并将其应用到零知识水印安全检测中。研究基于简单整数模运算和代数数域的完全同态加密方案，提高加密方案的执行效率，增强实用性；研究基于乘法同态加密的公开可验证零知识水印检测方案，防止第三方的欺骗；研究基于完全同态加密的非交互性零知识水印检测方案，提高水印检测的安全性。通过本研究不仅可从理论上建立一般性的零知识水印检测模型，且为水印和加密技术的应用提供新技术和新方法。

随着计算机技术和网络通信技术的迅猛发展，多媒体版权保护已经成为一个亟待解决的关键问题。数字水印技术作为信息安全领域的前沿方向，在多媒体信息版权保护和完整性认证方面得到了迅猛的发展。数字水印在进行验证或检测时一般需要出示私钥，而私钥一旦暴露，攻击者就能够伪造、修改和移去水印，水印检测协议一般需要可信第三方的参与，验证者在水印验证时也需要可信第三方的签名认证，这容易造成来自第三方的欺骗。本课题主要研究了零知识水印检测方案，文本文档数字水印嵌入和检测方案，云环境下数字水印的特征提取方法。本课题在对云环境下的数字水印提取方案研究时，发现对加密云数据进行语义检索是一个具有挑战性的关键难题，因此本课题在完成既定的任务同时，深入研究了云环境下密文检索技术。通过课题的研究，课题组发表学术论文36篇，其中sci收录15篇，培养硕士生7名。