PalmPhasor算法性能的理论分析

摘要

Security and privacy of templates are critical to the practical applications of palmprint systems. However, some objectives of biometric protection are difficult to meet at the same time due to the conflicts among them. As a cancelable palmprint coding algorithm to resolve the aforementioned conflicts, PalmPhasor is efficient and secure for palmprint authentication. In this paper, a complete analytical framework is proposed to systematically analyze the performance of PalmPhasor. The scenarios are categorized into four cases for the convenience of especial analysis. Some preliminaries, including auxiliary theorems, the distribution character of real and imaginary parts of Gabor filtered palmprint images, are provided to support the corresponding analysis. The theoretical analysis based on statistics and experimental results both confirm that PalmPhasor enhanced by multi-orientation score level fusion satisfactorily meets the four objectives of cancelable biometric at the same time even when users' tokens are stolen.%模板的安全性和隐私性是掌纹系统实际应用的关键问题,然而生物特征保护的多项指标通常相互冲突并且难以同时满足.作为解决上述冲突的一种可撤销掌纹编码算法,PalmPhasor实现了高效、安全的掌纹认证.建立了系统分析PalmPhasor性能的完整框架.为了便于具体分析,将情景分为4种情况,并且提供了支持相应分析的预备知识,包括辅助定理以及 Gabor 滤波掌纹图像实部和虚部分布特性.在统计学基础上建立的理论分析和实验结果均表明:即使在用户口令被盗的情况下,多方向分数级融合增强的PalmPhasor算法也可以同时有效地满足可撤销生物特征的4项指标.