基于视觉语义的云环境密文图像检索研究

云计算环境下密文图像检索是解决大规模图像存储管理困难与图像安全问题的关键，已得到各国政府、企业、研究人员的广泛关注和高度重视。如何使图像拥有者在云计算环境中加密存储自己的图像，又能让授权用户快速检索到自己需要的图像，是一个极具挑战性的课题。但这方面的研究仍处于起步阶段，有些关键问题还没得到有效解决。本项目拟从视觉感知的角度对云计算环境中密文图像检索所涉及的图像内容准确表征、高效索引构建等关键内容进行深入研究。主要研究内容包括：云计算环境中密文图像视觉语义形式化描述与模型构建、多特征融合的密文图像检索方法、基于感知鲁棒的安全索引构建、密文图像检索原型系统建立等。通过课题的研究，将在图像视觉语义特征提取、感知鲁棒哈希索引构建、多特征融合优化等方面完善和发展密文图像检索的相关理论和方法，为解决用户多需求查询、提高密文图像检索的精度和实用性提供新方法，为推动云环境下密文图像检索的发展提供技术支撑。

Encrypted image retrieval in cloud computing is a key technology to realize the massive images storage & management and images safety. It has risen extensive concern of governments, businesses and researchers. It is a challenging to fulfill the task of making the image owners to encrypt and store their images in cloud computing and allowing the authorized users to retrieve images quickly. The research on this task is still in its infancy, some key problems haven't been well solved. In the view of visual sensation, this program will dig deep into some key cruxes related to encrypted image retrieval such as the accurate characterization of image and effective index construction in cloud computing. The key research points in this project include: theories and model building of the visual semantic formalization of encrypted images, multiple features fusion encrypted image retrieval, sensation robust based safe index construction and encrypted image retrieve prototype system construction in cloud computing. This project will improve and develop the relevant theories and techniques of encrypted image retrieval on visual semantic feature extraction, robust hash index construction and multiple features fusion optimization and so on. It will provide new methods for multi-retrieval, better precision and application of encrypted image retrieval, and offer technique support to the development of encrypted image retrieval in cloud computing.