ID R&D breaks down new biometric spoof evaluation metric RIAPAR
The 2023 revision of the ISO/IEC 30107 standard for presentation attack detection gives weight to a new metric for evaluating false alarms in spoof detection, as explained in a blog post from ID R&D on the update.
The new update to the PAD standards’ Part 3 on testing and reporting introduces the concept of relative impostor attack presentation accept rate, or ‘RIAPAR,’ to the liveness standard.
RIAPAR is made up of two error rates; one for spoofs accepted as genuine biometric samples, and another for rejections of genuine biometric samples.
IAPAR, or impostor attack presentation match rate, contributes to RIAPAR by expressing the “likelihood of a successful attack on a complete biometric system.” RIAPAR, therefore, is a combination of IAPAR and FRR (false rejection rate). In this way, according to the blog post, it evaluates both the security and convenience of the system.
The standard has previously relied on attack presentation classification error rate (APCER) and bona fide presentation classification error rate (BPCER) to assess the effectiveness of PAD systems, which ID R&D argues “makes it difficult to assess how the technology will work in a real-world environment.”
The post breaks down how RIAPAR is calculated, and why it is important.
ID R&D also highlighted the financial impact of user friction introduced into customer onboarding processes from false positives in spoof detection in a recent case study.
While normally applied to face biometrics, ISO 30107 also applies to other modalities like voice.