保护图像数据隐私的托管计算研究

The widely-adopted mobile devices produce large amounts of image data every day and mobile users show an urgent desire of image sharing and searching. Due to the limited storage and computation resource of mobile devices, users often choose to outsource the complex processing like image search to cloud and image data has become one of the most important data on cloud. It is well known that image data include rich privacy information, thus massive outsourced image data leads to severe concern on privacy issues. Existing methods directly encrypt the image data which hurt their applications. Facing the challenges aroused by mobile big data applications, in this project we propose to study the privacy preserving outsourced image computing techniques and system. The applicants will conduct extensive research on the customized image privacy model based on correlation mining. We aim to support fine-grained privacy protection policy and significantly save the storage and computation cost. We study the outsourced image similarity computation scheme in privacy preserving manner, enabling authorized users to conduct content-based image search.

移动设备自身的有限资源难以满足大量图像数据的存储与计算需求，图像数据的托管已成为移动计算时代的必然趋势。然而图像数据包含丰富的个人隐私信息，不可靠的托管服务提供商对个人隐私带来了极大的威胁。简单的数据加密保护了数据隐私却牺牲了数据的可计算性，托管数据的隐私性和可计算性之间的矛盾已经成为了托管服务发展的瓶颈。.面对这一挑战，本课题拟对保护图像隐私的托管计算理论和系统进行研究。拟研究个性化的关联性图像隐私模型，提出对隐私数据的分离存储和细粒度访问控制机制，以显著降低隐私数据的存储和计算开销，并提升用户对个人数据的可控度。针对隐私数据的搜索计算，本课题拟提出可托管的保护隐私的图像相似度计算协议，从而同时保障托管数据的隐私性和可计算性，以填补该领域的空白。

随着移动计算时代的到来，图像数据的托管已经越来越普遍。然而，由于图像数据包含丰富的个人隐私信息，不可靠的托管服务提供商对个人隐私带来了极大的威胁。简单的数据加密保护了数据隐私却牺牲了数据的可计算性，托管数据的隐私性和可计算性之间的矛盾已经成为了托管服务发展的瓶颈。.面对这一挑战，本课题对保护图像隐私的托管计算的理论和技术进行了研究，构建并完善了图像隐私模型，研究了低开销的图像隐私托管存储与细粒度隐私保护机制，以及保护隐私的托管图像内容搜索方法，设计并实现了保护用户图像隐私的托管计算系统框架。此外，课题组还研究了与图像数据交易、分享相关的若干隐私保护问题。