基于高效隐秘汉明距离计算的安全人脸识别.pdfVIP

计算机工程与设计 Sept. 2016 2016 年 9 月 COMPUTER ENGINEERING AND DESIGN Vo l. 37 No.9 第 37 卷第9 期 基于高效隐秘汉明距离计算的安全人脸识别 刘卅1 ， 2 ，金鑫2+ ，赵耿2 ，李晓东2 ，陈迎亚2 ，郭魁2 (1.西安电子科技大学通信工程学院，陕西西安 710071; 2. 北京电子科技学院研究生系，北京 100070) 摘 要:为使人脸识别能够在云祷进行隐秘计算，实现一种能够安全计算的人脸识别系统。服务器端(云端)存储一纽嫌 疑人的人脸数据，客户竭(终端)获取人脸图像，在隐秘的情况下将其与服务器端的数据进行安全运算，判断客户端的人 脸是否与人脸库中的人脸匹配，若答案是肯定的，则返回给客户端匹配，不透露任何一方的信息。人脸识别采用人脸 特征块提取算法，安全计算采用同态加密和不经意传输算法，在此基础上提出一种高效隐秘汉明距离算法，解决在加密情 况下计算人脸向量汉明距离复杂度高的问题，程序测试时间明显减少。 关键词:人脸识别;安全;高效隐秘汉明距离;同态加密;不经意传输 (OT) 中圄法分类号: TP37 文献标识号:A 文章编号: 1000-7024 (2016) 09-2327-05 doi: 10. 16208/j. issn1000-7024. 2016. 09. 009 Security face recognition based on efficient secret hamming distance calculation l 2 2 LIU Yan • • JIN Xin抖. ZHAO Geni. LI Xiaσdoni. CHEN Ying-ya , GUO KuF (1. Scb∞1 of Telecommunications Engineering , Xidian University , Xi an 710071 , China; 2. Graduate Departrnent , Beijing Electronic Science and Technology Institute , Beijing 100070, China) At黯缸ad: To be able to carry out ∞vert face recognition ∞mputing in the clouds , a face r仅:ognition system for safe ∞mputing was achieved. Server (cloud) stored a set of suspects face database , the client (terminal) obtained face images in the ∞vert situation , and the images were proceeded using security operation with server side face database , and whether the client s face and the face in the library faces were matched was determined. If they were matched , the client would receive match without disclosing information of any party. The facial feature block extraction algorithm was used for face recognition , and Paillier homomorphic encryption algorithm an