基于属性基PIR的隐私保护云图像认证算法  

Privacy protection cloud graph authentication algorithm based on attribute-based PIR

作　　者：王超[1] 李晶[1] 张磊[1] WANG Chao;LI Jing;ZHANG Lei(College of Information Science&Electronic Technology,Jiamusi University,Jiamusi,Heilongjiang 154007,China)

机构地区：[1]佳木斯大学信息电子技术学院,黑龙江佳木斯154007

出　　处：《中国科技论文》2022年第11期1223-1229,共7页China Sciencepaper

基　　金：黑龙江省自然科学基金资助项目(LH2021F054);黑龙江省教育厅科技创新团队建设计划项目(2019-KYYWF-1335)。

摘　　要：针对云环境图像认证过程中潜在的泄露数据所有者及用户图像特征隐私的问题,提出了一种基于属性基隐私信息检索(private information retrieval,PIR)的隐私保护云图像认证算法,借助属性基PIR完成密态环境下的隐私图像认证。一方面,云服务器存储的由数据所有者提供的图像特征被加密,防止云服务器获取数据所有者隐私信息;另一方面,用户在认证时并不需要提供明文的数据特征,并且所提供的加密属性特征也经过泛化处理,最大程度地保护用户隐私。通过性能分析,在理论上证明了所提算法具有较好的隐私保护能力和算法执行效率;通过人脸和虹膜公共数据集进行了模拟实验测试,实验结果和成因分析证明了所提算法相比同类算法更为优越。In order to solve the potential problem of privacy of the owner in the process of graph authentication in cloud environment,a privacy protection algorithm for cloud graph authentication based on attribute-based private information retrieval(PIR)was proposed.The attribute-based PIR was used to complete the privacy graph authentication in the secret environment.On the one hand,the graph features stored by the cloud server and provided by the data owner were encrypted to prevent the cloud server from obtaining the privacy information of the data owner.On the other hand,users did not need to provide plaintext data characteristics during authentication,and the encryption attribute characteristics were also generalized to protect the user privacy to the greatest extent.Finally,through the performance analysis,it was theoretically proved that the proposed algorithm had better privacy protection ability and execution efficiency.Then experiments were carried out through the public data set of human faces and irises,further proving the superiority of the proposed algorithm.

关 键 词：隐私保护 云图像 图像认证 隐私信息检索 属性基加密 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP309.7[自动化与计算机技术—计算机科学与技术]