基于DCT排序特征加密的人脸图像安全检索系统

基于DCT排序特征加密的人脸图像安全检索系统 A Privacy Preserving Face Image Retrieval Scheme Based on DCT Order 尚月赟 SHANG Yueyun 1 first-author 李成蹊 LI Chengxi 2 叶登攀 YE Dengpan 2 中南民族大学数学与统计学院 中南民族大学数学与统计学院 School of Mathematics and Statistics,South-Central University for Nationalities School of Mathematics and Statistics,South-Central University for Nationalities 武汉大学计算机学院 武汉大学计算机学院 School of Computer,Wuhan University School of Computer,Wuhan University 在云环境下的内容存储与检索系统中,服务端不再是完全可信.使用常规加密方法,无法针对多媒体密文数据提供快速有效的内容检索服务.针对此问题,结合JPEG编码特点本文提出并实现了一种基于Logistic混沌加密和基于DCT排序的隐私保护内容检索系统.该系统可以实现JPEG图像的视觉加密,同时并不影响用于检索的特征和效率,从而使得服务端可以对密图进行内容检索,并且算法能够忍受一定程度的噪声和压缩.其性能特点得到了实验验证. In the cloud content storage and retrieval systems,the server can not be fully trusted.If using conventional encryption method to protect cloud data,the retrieval efficiency to the multimedia content ciphertext data will be restricted to a great extend.In this paper,aprivacy preserving image retrieval scheme based on DCT coefficient order and Logistic Map was proposed.The JPEG image can be visually scrambled without changing the retrieval feature and efficiency,and the function of the content retrieval in encrypted domain can be realized.And it can resist some noises and compression.All the above performance were verified by the experiments. 隐私保护 图像检索 DCT排序 Logistic映射 privacy preserving image retrieval DCT order Logistic map 10.14188/j.1671-8836.2014.06.014 TP391.41 国家自然科学基金（61272453）;湖北省自然科学基金（2014CFB379）资助项目 2014-06-01 2021-04-01 6 60