The field of biometrics has rapidly grown over the past few years with new companies entering the market every month. This massive growth caused by the greater demand for biometric services makes it crucial for customers and users to ascertain the quality of the technologies used to be certified and assessed under the strictest and most updated standards and regulations.
In this sense, Veridas has recently become the only company to be included in NIST’s evaluations for facial recognition (both 1:1 and 1:N) and speaker recognition (1:1). In addition, Veridas has passed ISO 30.107 iBeta Level 1 PAD (Presentation Attack Detection) test for its Liveness Detection solution and also holds ISO 27.001 on the Security of Information Systems accredited by Bureau Veritas.
What do these certifications mean?
- NIST 1:1 Facial Recognition: This report evaluates the performance of a biometric system on comparing different pictures of the same individual. In April 2021, Veridas achieved a false negative rate (mistakenly rejecting someone) of 2.86% and a false positive rate (mistakenly accepting someone) of 0.001% in the wild category, where pictures are captured in a real-life environment (different angles, shades, etc).
- NIST 1:N Facial Recognition: This consists of identifying an individual among a database of 1.6 million registered users. In April 2021, Veridas engine obtained a false negative rate of 1.4% for a false acceptance rate of 1%.
- NIST 1:1 Speaker Recognition: The basic task in this challenge is determining whether a specified target speaker is speaking during a given segment of speech. Veridas engine is currently under this evaluation. The leaderboard can be checked here.
- ISO 30.107 iBeta Level 1 PAD: This test evaluates the performance of a liveness detection solution against spoofing attacks such as high-quality photos or 2D and 3D paper masks. Veridas technology passed the iBeta test with an Attack Presentation Classification Error Rate (spoofing detection) of 0% for the total 1800 attacks that were attempted.
- ISO 27.001: This certification is widely known for providing companies and other entities with requirements for an information security management system (ISMS). Veridas has been voluntarily evaluated to acquire this certification to ensure compliance with GDPR and NIS Regulations among others.
With these external evaluations, Veridas submits its technologies to the highest evaluation standards in terms of transparency, quality of the algorithms and low gender and racial bias, aligning with the European Proposal for a Regulation laying down harmonised rules on artificial intelligence (Artificial Intelligence Act), proposed by the European Commission on the 21st of April.
NIST, the institution of reference in the field of biometrics
The National Institute of Standards and Technology in the US is the most relevant institution in terms of biometric technologies’ standardization. NIST carries out periodic evaluations of biometric engines submitted by different providers to assess and provide unbiased accuracy and performance rates in different environments.
NIST participation is completely free of charge and open to any vendor or institution.
Veridas has been submitting its technologies to NIST’s evaluations since 2018 and continues to do so with the strong belief that constant scrutiny is the only way to maintain its position as a global leader in the biometric solutions market.
Results shown from NIST do not constitute an endorsement of any particular system, product, service, or company by NIST.
Biometric technologies available for everyone
Veridas solutions are already improving customer engagements in several companies around the world from different sectors such as banking, insurance, telecommunications or public administrations. In particular, Veridas facial biometric engine has been used for fraud detection in the financial sector where it enables the identification of duplicate identities comparing 1 photo versus a 1.000.000 database in less than a second, bringing a whole new dimension to the world of fraud prevention.