signzy

API Marketplace

downArrow
Logo
Responsive
Decorative line

FAR / FRR (biometric error rates)

Overview

False Acceptance Rate (FAR) and False Rejection Rate (FRR) are key performance metrics used to evaluatebiometric authentication systems.. FAR measures how often an unauthorized person is incorrectly accepted as legitimate, while FRR measures how often a legitimate user is wrongly rejected. Together, they reflect the trade-off between security and usability in biometric systems.
For banks, fintechs, healthcare providers, government agencies, and telecom operators, balancing FAR and FRR is critical. A high FAR increases fraud risk, while a high FRR frustrates legitimate users and disrupts operations. Biometric technologies such as facial recognition, fingerprint scanning, and voice authentication must be fine-tuned to achieve an optimal balance, often represented by the Equal Error Rate (EER) where FAR and FRR intersect. Monitoring these metrics ensures biometric systems remain secure, accurate, and user-friendly while complying with regulatory and data protection requirements.

FAQ

Why do FAR and FRR matter?

They quantify biometric risk and friction. Lower FAR reduces unauthorized access; lower FRR reduces user lockouts. The “right” balance depends on product risk, regulatory expectations, and customer experience goals.

What is EER and when to use it?

Equal Error Rate is the point where FAR equals FRR, useful for comparing models in labs. In production, choose thresholds aligned to use case risk, not the EER point.

How to measure in practice?

Use representative datasets, include challenging demographics and environments, and report confidence intervals. Re-test after model updates or camera changes to ensure consistency and fairness over time.

How do attacks affect rates?

PAD-tested scenarios matter: if spoof prevalence rises, effective FAR can increase. Incorporate attack-aware testing, liveness, and continuous monitoring to keep risk within appetite.