ROC Posts Sub-One-Percent Miss Rates on NIST FRIFTE Class B Slap Fingerprint Evaluation

ROC, the US biometrics company formerly known as Rank One Computing, has posted strong results on the National Institute of Standards and Technology’s Friction Ridge Image and Features Technology Evaluation Exemplar One-to-Many track, the standing third-party benchmark for fingerprint search algorithms. On the FRIFTE E1N Class B test set, which uses slap fingerprints against a three-million-subject enrollment database, ROC’s submission produced rank-one miss rates of 0.24 percent for left slap, 0.36 percent for right slap, and 0.00 percent when both slaps were combined as the probe.

FRIFTE E1N is the successor to NIST’s FpVTE 2012 evaluation and tests how well an algorithm identifies a person from one or more fingerprint images against a large enrollment database. Class B specifically uses slap captures, the rolled-flat impressions of four fingers at a time that are taken at booking stations, border posts, and civil enrollment kiosks, against a reference database of complete ten-finger identification flats. The configuration is the operational match for national identity, border control, and law enforcement deployments that rely on slap captures rather than rolled or latent prints.

The NIST report card for ROC’s roc+0002 submission, dated 5 January 2026, sets out the algorithm’s measured error rates across the Class B sub-tests. At a False Positive Identification Rate of 0.01, ROC’s submission produced False Negative Identification Rates of 0.0033 for left slap, 0.0040 for right slap, and 0.0002 for both the combined left-and-right slap probe and the complete identification flat probe. Rank-one accuracy on the cumulative match characteristic measurement was 0.9995 for the identification flat probe and 1.0000 for the combined slap probe.

The result extends ROC’s recent run on the NIST evaluations. The company has previously claimed a top global ranking on NIST’s Evaluation of Latent Fingerprint Technologies and posted top combined accuracy and efficiency on earlier NIST testing. ROC positions the Class B result for national ID, border control, and large-scale civil enrollment deployments, where slap captures are the standard input modality.

NIST notes in the report card that the information about the submission was provided by ROC and not verified by NIST, and that the participation identifier is roc+0002. The underlying test dataset, FpVTE 2012, was used to maintain comparability with earlier NIST fingerprint evaluations.

Sources: NIST FRIFTE E1N program page

–

By the ID Tech Editorial Team