彩色图像隐私部分自动标定及加密  

Automatic calibration and encryption of privacy part of color image

作　　者：张赛男 李千目[1] Zhang Sainan;Li Qianmu(School of Computer Science and Engineering,Nanjing University of Science and Technology,Nanjing 210094,China)

机构地区：[1]南京理工大学计算机科学与工程学院,江苏南京210094

出　　处：《南京理工大学学报》2022年第4期395-405,共11页Journal of Nanjing University of Science and Technology

基　　金：国家重点研发计划(2020YFB1804604);工业互联网创新发展工程(TC190A3X8-9);江苏省重点研发计划(BE2016904);装备预先研究项目(41412***705)。

摘　　要：为了进一步研究图像隐私加密保护,该文构建了一种安全可靠有效的基于人脸检测改进的彩色图像隐私部分自动标定及加密算法。利用先进的人脸检测模型初步检测图像中的人脸隐私信息;合并扩展图像隐私区域;基于logistic-sine-cosine映射设计置乱扩散算法来加密图像隐私部分。该文提出的自动检测、合并标定图像隐私区域的算法能够有效覆盖并加密保护图像隐私信息。加密算法的密钥空间大,密钥灵敏性高,能够抵抗相关性分析、直方图分析、信息熵等统计分析及差分攻击等,具有良好的混淆和扩散性。In order to further study the image privacy encryption protection,this paper constructs a safe,reliable and effective color image privacy part automatic calibration and encryption algorithm based on improved face detection.First,the advanced face detection model is used to initially detect the face privacy information in the image.Then the privacy area of the image is improved and merged.Finally,a scrambling diffusion algorithm is designed based on the logistic-sine-cosine mapping to encrypt the privacy part of the image.The algorithm proposed in this paper can automatically detect,merge and demarcate the privacy area of the image,which can effectively cover and encrypt the privacy information of the image.The encryption algorithm has a large key space and high key sensitivity.It can resist statistical analysis such as correlation analysis,histogram analysis,information entropy,and differential attacks,which has good confusion and diffusion.

关 键 词：部分加密 隐私保护 自动标定 人脸检测 彩色图像 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]