局部置乱结合双随机相位编码的双虹膜身份模板保护方法

Double iris identity template protection method combining partial scrambling and double random phase coding

针对基于不可逆变换的虹膜模板保护在进行虹膜识别时面临特征模板泄露的问题,提出一种结合局部置乱和双随机相位编码双虹膜身份模板保护方法。该方法首先采用log-Gabor变换提取用户虹膜特征,然后设定置乱次数以及变换矩阵参数,利用Arnold变换对虹膜特征矩阵进行无重叠分块置乱,对原始特征进行第一重加密保护。再在分数傅里叶变换域应用双随机相位编码方法构建加密特征向量,采用用户设定掩膜M1和同一用户的另一只虹膜图像掩膜M2对置乱矩阵进行第二重加密保护。最后依据汉明距离对加密模板进行匹配识别。为验证算法性能,采用CASIA-IrisV3-Interval、MMU-V1以及CASIA-IrisV4-Lamp虹膜图库的图像进行测试。结果表明,该方法识别正确率能够达到98.73%,可生成模板的总数量为522,并且恢复原始数据需要尝试31452×624×42次,能够满足生物特征模板保护的不可逆性、可撤销性和不可链接性标准。

Aiming at the problem that iris template protection based on irreversible transformation is faced with feature template leakage in iris recognition, this thesis proposes a method of double iris identity template protection method combining partial scrambling and double random phase encoding. Firstly, the iris feature of the user is extracted with log-Gabor transform, and then the scrambling times and transformation matrix parameters are set. The non-overlapping block scrambling of iris feature matrix is carried out using Arnold transformation, and the first encryption protection of the original features is carried out. Then, double random phase coding method is used to construct the encryption feature vector in fractional Fourier transform domain. The user set mask M1 and the same user′s other iris image mask M2 are adopted to perform the second encryption protection of the scrambling matrix. Finally, the Hamming distance is used to match and identify the encryption template. In order to verify the performance of the algorithm, the iris images in the CASIA-IrisV3-Interval, MMU-V1 and CASIA-IrisV4-Lamp iris libraries were used for testing. The results show that with the proposed method the recognition accuracy can reach 98.73%, the total number of the generated templates is 522, and the recovery of the original data needs 31452×624×42 attempts, which can meet the criteria of irreversibility, revocability and unlinkability of biometric template protection.