|
|||||||||
|
| You are here: CESG > Policy and Technologies > BWG > Management Summaries > | ||||||||||||
|
Best Practices of Testing and Reporting Biometric Device Performance - MS11
The scientific, technical testing of biometric devices, including the measurement of error and throughput rates, has a 25 year history. Over this period, results of dozens of carefully controlled tests have been published. However, as there has been no commonly agreed upon method for conducting or reporting the tests, it has been impossible to compare and interpret results across time, across laboratories, or across devices. In 1999, the BWG began the task of developing a normative, general approach to evaluating and reporting biometric device performance - including the definition of basic metrics and the specification of minimum reporting elements - and documenting historical “lessons learned”. Detail The starting point for the BWG “Best Practices” was the highly refined and respected test protocol from the (U.S.) National Institute of Standards and Technology (NIST) Speaker Verification Test Program. This protocol is a legacy document from nearly 40 years of US government speaker verification research. Issue 1.0 of the BWG “Best Practices” was an attempt to directly translate the NIST protocol to more general biometric device testing. This issue was widely circulated in the testing community and many comments were received. These comments, and the experiences of the UK National Physical Laboratory (NPL) in conducting their own detailed test in 2000, were incorporated into an Issue 2.0 of “Best Practices”, released in 2002. We are now working on the next issue, which will incorporate closed-set testing and non-parametric reporting protocols. “Best Practices” is not intended to be dogmatic or doctrinaire, but rather simply a documentation of collected, current wisdom. We understand that the vagaries and challenges of testing the performance of human populations with emerging biometric technologies will require flexibility in its application. We expect this document to undergo continuing revision, as we gain experiences and broaden our understanding of the challenges of testing these technologies at the human/machine interface. For further information see: A copy of the “Best Practices (pdf)” can be downloaded from the BWG web pages Back to Management Summary Index The UK Biometric Working Group, managed by CESG, supports the UK government and provides advice and information about the implementation and use of biometric authentication systems. For further details telephone +44 (0) 1242 221491 extension 34124 |
|||||||||||
 |
|||
|