Defining classification thresholds in template protection systems
    1.
    发明授权
    Defining classification thresholds in template protection systems 有权
    定义模板保护系统中的分类阈值

    公开(公告)号:US08412950B2

    公开(公告)日:2013-04-02

    申请号:US12808579

    申请日:2008-12-17

    IPC分类号: H04L29/06

    摘要: A method for configuring a biometric template protected authentication system wherein the desired classification threshold is first selected to optimize the trade-off between a false accept (FAR) and a false non match or reject rate (FRR) of the system, and then an error correcting code (ECC) used in the authentication process is chosen such that the number of errors which can be corrected is equal to or greater than the selected classification threshold. During authentication, the number of errors in a first codeword derived from biometric data associated with a physical object is determined and used in the decision process to accept or reject authentication.

    摘要翻译: 一种用于配置生物特征模板保护认证系统的方法,其中首先选择期望的分类阈值以优化系统的错误接受(FAR)和错误不匹配或拒绝率(FRR)之间的折衷,然后是错误 选择在认证处理中使用的校正码(ECC),使得可以校正的错误的数量等于或大于所选择的分类阈值。 在认证期间,在与物理对象相关联的生物特征数据中导出的第一码字中的错误数量被确定并在决策过程中用于接受或拒绝认证。