摘要
回顾总结生物特征识别系统算法性能指标和算法评测方面的最新研究进展,详细介绍相应算法评价指标的定义、实现过程及意义,不仅总结现有的基本指标量并将其一般化公式化,更具体描述P-R曲线、ROC曲线、DET曲线、CMC曲线、等错误率EER和类内类间差异评价指标d'等综合性指标的绘制实现过程及相关属性意义。最后,还对生物特征识别系统相关的其他指标进行简单介绍。
Review and summarize the latest research progress of algorithm performance indicators and algorithm evaluation of the biometric recognition system,introduce the definitions,implementation processes and meanings of corresponding algorithm evaluation indicators in detail.This paper not only summarizes the existing basic indicators and formulates them in general,but also describes the drawing implementation processes and related attribute meanings of comprehensive indicators such as P-R curve,ROC curve,DET curve,CMC curve,EER and the difference of intra-class and inter-class d'.Finally,we also introduce other indicators of the biometric recognition system.
作者
高乾坤
张鹏
徐平
徐骁
张照星
陈洁
施一琳
GAO Qian-kun;ZHANG Peng;XU Ping;XU Xiao;ZHANG Zhao-xing;CHEN Jie;SHI Yi-lin(Research Center of Census Administration,M.P.S.PRC,Beijing 100070)
出处
《现代计算机》
2020年第28期8-11,46,共5页
Modern Computer
基金
公安部科技计划重点项目(No.2019JSYJA07)。
关键词
生物特征识别
性能指标
算法评测
Biometric Recognition
Performance Indicator
Algorithm Evaluation