云计算下的加密域多媒体水印与模式匹配

In recently, signal processing in the encrypted domain has become one attracting domain in research field of information security. By combining encryption algorithms and signal processing techniques, signal processing in the encrypted domain provides a promising solution protect multimedia security and users' privacies, satisfying the requirements of data security in distributed network, e.g., cloud computing. In the background of big data, it requires more efficient data encryption algorithms and processing methods for signal processing in the encrypted domain. This project mainly studies digital images and videos in big data for the purpose of protecting multimedia security and users' privacies, by building statistics models, designing encryption algorithms and proposing signal processing methods. The research contents are mainly focus on: 1) piracy preserving face recognition based on encrypted domain discrete wavelet transform; 2) homomorphic encryption algorithm for digital image database in big data; 3) secure image watermarking scheme against encrypted domain watermark attacks. Our research achievements can optimize the theory and methods of signal processing in the encrypted domain, and provide solid technical supports for the applications in this the field.

近年来，加密域信号处理已经成为了信息安全中一个备受关注的领域。运用加密算法和信号处理技术，加密域信号处理能为多媒体安全和用户隐私保护提供了新颖且有效的解决方案，满足在分布式网络（如云计算）环境下对数据安全性的需求。在大数据背景下，加密域信号处理需要效率更高的数据加密算法和处理算法。本项目主要以大数据下的数字图像/视频为研究对象，通过建立统计模型，设计加密算法和提出信号处理技术来保护实多媒体安全和用户隐私。主要研究内容包括：1)基于加密域小波变换的保护用户隐私的人脸识别方案；2)大数据库下图像库的同态加密算法；3)抵抗加密域攻击的安全图像水印方案。本项目的研究能进一步完善加密域信号处理领域的理论和方法，并为该领域的实际应用提供可靠的技术支撑。

如今，数字媒体（主要研究数字图像及视频）已经成为人们获取及传递信息的最主要载体之一。然而，随着云储存和云计算的普及，数据盗取和隐私泄露的案例越来越多，这将不可避免地对我们个人隐私、法律取证、科学发展乃至国家安全带来一定的负面影响。加密域信号处理具有十分重要的研究及实际意义。本项目主要围绕着媒体安全若干关键问题展开研究，主要研究成果包括了: （1）基于混合变换域的加密图像水印嵌入；（2）加密H.264/AVC视频上的运动检测和跟踪；（3）高效率的加密域图像变换及其并行化。