基于人脸特征的可搜索加密方案  

SEARCHABLE ENCRYPTION SCHEME BASED ON FACE FEATURES

作　　者：徐慧华 杨雄 Xu Huihua;Yang Xiong(Department of Economics and Law,Concord University College Fujian Normal University,Fuzhou 350117,Fujian,China;Department of Computer Engineering,Zhicheng College of Fuzhou University,Fuzhou 350002,Fujian,China)

机构地区：[1]福建师范大学协和学院经济与法学系,福建福州350117 [2]福州大学至诚学院计算机工程系,福建福州350002

出　　处：《计算机应用与软件》2022年第11期304-309,共6页Computer Applications and Software

基　　金：2018年福建省高校杰出青年科研人才培育计划项目;2019年福州大学至诚学院省基金培育项目(KJ2019004)。

摘　　要：为防止人脸特征泄漏,设计基于全同态加密的人脸特征加密方案,确保网络传输和数据库中保存的是人脸特征的同态密文。该方案可通过人脸特征同态密文的计算完成人脸匹配。实现部分借助微软的开源简单加密算法函数库,并进一步探索一种批处理和近似最近邻搜索的方法,提升加密域中人脸匹配的效率。基于人脸基准数据库(LFW)的测试结果表明,基于全同态的人脸特征加密和近似最近邻的人脸搜索匹配是可行的,能够满足实际应用场景的需求。To prevent the leakage of facial features,a facial feature encryption scheme based on fully homomorphic encryption is designed to ensure that the homomorphic ciphertext of facial features is stored in the network transmission and database.In this scheme,face matching could be accomplished by calculating homomorphic ciphertext of face features.The implementation part used the open source simple encryption algorithm function library of Microsoft,and further explored a batch processing and approximate nearest neighbor search method to improve the efficiency of face matching in the encrypted domain.The test results based on the face reference database(LFW)show that face feature encryption based on full homomorphism and face search matching with approximate nearest neighbors are feasible and can meet the needs of practical application scenarios.

关 键 词：全同态加密 人脸特征 最近邻搜索 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]