iBeta is accredited by NIST NVLAP as an Independent Test Lab and the ISO/IEC 19795-2 standard is included in our scope of accreditation. iBeta conducts biometric performance testing in accordance with the ISO/IEC 19795-2 standard, providing both scenario and technology evaluation results. We present the following information to clarify our procedures and standards.  

The iBeta 19795-2 test procedure typically involves a scenario test, which includes a live data collection, giving clients feedback and insight as to how their system performs under real-world conditions. 

Prior to test start, iBeta determines the client’s needs based on the desired FAR/FRR metrics and assesses whether the associated software and hardware can support the scenario test, which includes test subjects performing enrollment and verification transactions.  To pass iBeta’s ISO/IEC 19795-2 performance test, the client’s system must meet the error metrics defined in the table below. Additionally, iBeta’s standard test method can report results down to 0.001% for FAR and 0.25% for FRR at a 95% confidence interval. This rigorous testing can provide your product with a competitive advantage. 

  iBeta ISO/IEC 19795 

Requirements 
False Accept Rate (FAR)  0.01% 
False Reject Rate (FRR)  5% 

95% Confidence Interval 

These are the statistical capabilities of iBeta’s standard 19795-2 testing procedures, which can be adjusted by client request, depending on need. Including the above, iBeta reports on the following performance metrics:  

Abbreviation  FTER  FTAR  FNMR  FMR  FAR  FRR  GFRR  GFAR 
Metric  Failure to Enroll Rate  Failure to Acquire Rate  False Non-Match Rate  False Match Rate  False Accept Rate  False Reject Rate  Generalized False Reject Rate  Generalized False Accept Rate 

Note that iBeta also conducts performance testing as integral parts of its FIDO Biometric Component Certification testing, Mastercard Biometric Certification, and Android Biometric Security Test Protocols to the relevant required metrics. 

iBeta posts, with written client consent, the performance testing confirmation letters that provide testing results. These letters specify the type of PAD testing and the configuration of the client product that was tested.  To fully assess the client product, the audience of these letters should understand the test method used to assess conformance with the ISO 19795-2 Standard. 

ISO/IEC 19795-10 – Performance Testing with Demographic Bias Evaluation 

Demographic bias is a known issue within biometric recognition systems and can provide inaccurate results for systems not appropriately evaluated for bias. The ISO/IEC 19795-2 standard does not address issues of bias, because it does not address the comparisons of results between demographic groups. Considering this, iBeta has pursued and received an ISO/IEC 19795-10 accreditation, allowing for performance testing procedures that focus directly on potential demographic bias in biometric systems.  

Statistical methods are used to compare the confidence intervals of the FAR/FRR rates across gender, age, and Monk skin tone to determine if there is potential bias in the system towards any demographic categories.  

By submitting a system for demographic bias testing, client can examine their systems’ performance robustness within demographic groups, address any areas of concern, and in turn assure future partners that those systems meet any relevant regulatory guidelines.  

In pursuing either iBeta’s ISO/IEC 19795-2 or ISO/IEC 19795-10 test, the client has the option to include compliance with the other standard in the test. 

DateVendorStandardProduct TestedVersion(s) TestedFRRPDF Link
08/16/24Vietnam Payment Solution (VNPAY)19795-2VNPay Face EKYC1.0 (34)5% FRRPDF