As presentation attack detection becomes more well known as an important element in any biometric system being used for fraud prevention or security, ID R&D has published a backgrounder to help customers and implementers understand how the performance of this technology mis measured.

Like accuracy in biometric systems, the effectiveness of presentation attack detection is sometimes misunderstood, and presented for instance as a single percentage expressing ‘when it works.’

ID R&D’s white paper introduces the concepts of attack presentation classification error rate (APCER) and bona fide presentation classification error rate (BPCER), and how they are used.

‘Presentation Attach Detection: Measuring Performance with APCER and BPCER’ is available for free download with registration. It draws an analogy to error measurements for biometric algorithms; false match rate and false non-match rate (FMR and FNMR), as well as false acceptance rate and false rejection rate (FAR and FRR).

“They are similar, but there are reasons why we have three sets of error rates,” the document explains.

Sections on statistical hypothesis testing and error rates in biometrics are followed by one specifically on error rates in presentation attacks. The concept of DET (“detection error tradeoff”) curves is introduced, and a glossary provided to help readers follow the technical points made.

ID R&D offers PAD for voice and face biometric modalities, both through ‘passive’ checks, which the company argued for in a previous white paper.

The company also worked with BixeLab to develop bias testing for liveness detection, with results unveiled earlier this year.

biometric liveness detection  |  biometrics  |  ID R&D  |  passive facial liveness  |  presentation attack detection  |  white paper

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Copyright © 2022 Biometrics Research Group, Inc. All Rights Reserved.