TECH5 said Monday that a new NIST FRIF TE 1:N evaluation shows its latest fingerprint matching algorithm delivering unusually fast identification speeds at large civil identity scale. The new algorithm ranked among leading participants on accuracy while posting mean search times in the millisecond range on a five million record gallery.
NIST’s Class B dataset results show a mean search time of 0.0 seconds, which TECH5 said most closely reflects real-world national ID and civil registry deployments. The next-fastest algorithm in the evaluation recorded about 3.7 seconds per search, while other participating systems ranged from roughly five seconds to more than two minutes on the same database sizes.
The performance gap matters most in the kinds of national-scale deployments that are expanding across the digital identity market. Somalia recently began enforcing a biometric national ID requirement for domestic flights, and ongoing work to secure Estonian remote biometrics reflects the throughput pressure on identity systems as databases and transaction volumes grow.
TECH5 said the tested engine is already deployed in its T5-OmniMatch ABIS platform, both as a standalone system and as part of its wider digital public infrastructure offering. The company framed the benchmark as a cost and throughput issue as much as an accuracy issue, arguing that faster 1:N searches can lower infrastructure requirements and improve response times across very large databases.
The result also highlights how benchmark discussions are shifting from pure false-match rates to operational performance. For public agencies and system integrators, the question is not only whether an algorithm is accurate enough, but whether it can search very large galleries quickly enough to support enrollment, investigation and service-delivery workloads without creating infrastructure bottlenecks.
Sources: TECH5, NIST FRIF TE 1:N results
–
By the ID Tech Editorial Team
