Identy.io’s latest fingerprint-matching algorithm recorded a 19 percent lower error rate than the company’s February entry in NIST’s Friction Ridge Image and Features Technology Evaluation, or FRIF TE E1N, Biometric Update reported on October 22.
FRIF TE E1N evaluates the matching algorithms used at the core of automated biometric identification systems. In the reported Class C comparison of plain fingerprints against rolled fingerprints, Identy recorded a false non-identification rate of 0.0008 at a false positive identification rate of 0.0001 or lower, according to the report.
The company’s Identy+0002 algorithm also recorded 99.99 percent identification accuracy for slap and flat fingerprints against a Class B database containing 3 million subjects. Searches at that scale had a reported median time of 2.9 seconds, Biometric Update reported. These figures are results attributed to the NIST evaluation; the available report does not describe the sensors, image conditions or demographic composition behind the comparisons.
The submission is one of six recent additions to the benchmark. The report also lists results from Algorithms Inc., Beijing Hisign, Miaxis, Griaule and NEC, but the comparisons use different classes and metrics, so the figures should not be treated as a simple overall vendor ranking. The report provides the individual comparison results.
Identy’s company profile says its automated biometric identification system supports facial, fingerprint and iris matching and is currently deployed with airports and existing border infrastructure. That description provides deployment context, but it does not independently validate the new benchmark results or document the operational conditions of those deployments.
For buyers, the reported improvement is evidence of progress in the FRIF TE E1N evaluation, not by itself proof of production readiness. The fetched material does not provide pricing, integration requirements, error-handling procedures or privacy safeguards for adopting the system.