基于局部排序的双虹膜模板保护方法

A Method for Protecting Double Iris Template Based on Partial Sorting

针对不可逆变换虹膜模板保护方法在虹膜识别时的特征模板泄露问题,提出一种基于局部排序的双虹膜模板保护方法。方法采用log-Gabor变换提取用户的左右虹膜特征,将左右虹膜特征数据进行异或,对左虹膜特征数据与异或所得结果分别加以处理。在处理过程中设置块参数,分组处理所分的块数,将排序数值换算为二进制序列作为最终的加密模板,从而依据汉明距离对加密模板进行匹配识别。采用多套虹膜图库进行测试,验证算法性能,实验结果表明,识别正确率能够达到90.55%,满足生物特征模板保护的不可逆性。

Aiming at the problem of feature template leakage in iris recognition by irreversible transformation iris template protection, a double iris template protection method based on partial sorting is proposed. The log-Gabor transform is used to extract the users’ left and right iris features, and the left iris feature data are XORed, and the left iris feature data and XOR results are processed respectively. In the process of processing, block parameters are set, the number of blocks is divided into groups, and the sorting value is converted into binary sequence as the final encryption template, thus matching and identifying the encryption template according to Hamming distance. Several sets of iris image libraries are used to test the performance of the algorithm. The experimental results show that the recognition accuracy can reach 90.55%, which meets the irreversibility of biometric template protection.