面向虹膜模板的复合映射加密算法  

Composite mapping encryption algorithm for iris template

作　　者：李美珊[1] 刘越[1] 张磊[1] 薛佳楣[1] LI Meishan;LIU Yue;ZHANG Lei;XUE Jiamei(College of Information Science and 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》2021年第3期288-294,共7页China Sciencepaper

基　　金：中国博士后科学基金资助项目(2019M661260);黑龙江省自然科学基金资助项目(YQ2019F018);黑龙江省省属本科高校基本科研业务费科研项目(2018-KYYWF-0938);佳木斯大学学科团队项目(JDXKTD-2019008);佳木斯大学自然科学科研项目(12Z1201520)。

摘　　要：针对远程生物认证中虹膜模板图像既需要加密保存又要保证可恢复性的特点,在使用卷积神经网络(convolutional neural networks,CNN)对特征像素矩阵提取的基础上,提出一种复合映射的虹膜模板加密算法。将CNN提取的像素特征矩阵分组,并对不同分组分别使用阿诺德映射改变权值矩阵实现矩阵扰乱。使用分段线性混沌映射隐藏邻近相关像素,完成虹膜图像模板加密。由于映射分别针对特定分组的特征像素矩阵,具有更好的原始图像可恢复性。实验结果表明,所提出的算法对远程认证的虹膜模板图像具有很好的加密性能和较高的算法加/解密执行效率。In remote biometric authentication,the iris template image not only needs to be encrypted and preserved,but also can be recovered.Based on the feature pixel matrix extraction using convolutional neural networks(CNN),an iris template encryption algorithm based on composite mapping was proposed.The pixel feature matrix extracted by CNN was grouped,and Arnold map was used to change the weight matrix for different groups to achieve matrix perturbation.Using piecewise linear chaotic map to hide the adjacent pixels,the iris image template encryption was accomplished.Because the mapping is aimed at the characteristic pixel matrix of a specific group,the original image can be recovered better.The experimental results show that the proposed algorithm has good encryption performance and high encryption/decryption efficiency for remote authentication iris template images.

关 键 词：虹膜模板 图像加密 卷积神经网络 阿诺德映射 分段线性混沌映射 

分 类 号：TP309.2[自动化与计算机技术—计算机系统结构]