Innovatrics Achieves 6x Fewer Misses in NIST FRIF TE E1N Class A Fingerprint Identification Test

NIST’s Friction Ridge Image and Features (FRIF) Technology Evaluation (TE) Exemplar One-to-Many (FRIF TE E1N) benchmark measures how accurately fingerprint identification algorithms find the right person in very large databases, alongside speed and template size

Innovatrics has achieved a sixfold reduction in misses in the latest NIST FRIF TE E1N evaluation, specifically in the Class A test using both index fingers. This test reflects a common setup in large-scale fingerprint identification systems, where accuracy must be maintained at strict false-match limits.

Against an enrollment of roughly 1.6 million subjects, Innovatrics reduces the miss rate in the Class A two-index-finger test from 0.003 to 0.0005 at the same false-positive identification rate of FPIR ≤ 0.001. This means six times fewer missed matches without increasing false matches.

By lowering the miss rate at a fixed FPIR, the updated results show improved matching performance without trading accuracy for risk. This matters most in large databases, where even small percentage changes can translate into thousands of avoided misses.

Efficiency and Template Size

NIST also reports performance metrics beyond accuracy. In the same evaluation, Innovatrics shows a median template creation time of 0.3 seconds for a single index finger and 0.6 seconds for both index fingers.

Template size is similarly compact. The median size for a single index finger is 2.2 kB, the smallest among the evaluated submissions. Smaller templates reduce storage, memory use, and data transfer requirements, which becomes increasingly important as systems scale.

What This Means for Partners

For deployments, these improvements translate into fewer missed hits while maintaining a low false-match setting. These help reduce operator workload, repeat captures, and infrastructure pressure in large-scale systems such as national ID, border, and law enforcement.

Results shown from NIST do not constitute an endorsement of any particular system, product, service, or company by the U.S. Government.