Innovatrics Tops the NIST ELFT Benchmark (December 2023)
Innovatrics’ latest submission to the NIST ELFT benchmark confirms its latent fingerprint algorithm is the most accurate in the world, outperforming all competition in rank-based search.
NIST Evaluation of Latent Fingerprint Technologies (ELFT) consists of evaluations of the accuracy of latent matching using features marked by experienced human latent fingerprint examiners. This test aims to evaluate the current state of the art in latent feature-based matching by comparing the accuracy of searches using images alone with searches using different feature sets.
Innovatrics has the most accurate algorithm with a rank-1 hit rate of 93.3%, further strengthening its position as a leading provider of cutting-edge biometric solutions. According to the NIST ELFT Innovatrics nutrition report, Innovatrics is also top ranked in accuracy across all categories.
Hit rate is the most important metric derived from FNIR (false negative identification rate) and the resulting value is the percentage of times that the correct subject appears at or above a rank in the candidate list. This is nowadays the most widely used method by fingerprint laboratories across the world where the results are reviewed manually.
September 2023 Results
In September 2023, Innovatrics outperformed all the competition in identifying latent fingerprints as measured by the NIST ELFT test.
Innovatrics reached the top with a rank-5 hit rate of 95.1%, surpassing the competition as seen in the chart from the NIST ELFT nutrition report below and further strengthening its position as a leading provider of cutting-edge biometric solutions.
Innovatrics has scored exceptionally well in the so-called “lights-out” searches, a category where a threshold is used to filter out more than 90% of non-matching candidates. A higher threshold likely corresponds with fewer subjects for manual review, which saves the operator some time. With a hit rate of 82.4%, Innovatrics is ranked 2nd in a category where the Rank-100 Hit Rate was measured at FPIR = 10%. This is also NIST’s preferred accuracy metric because operating costs grow with the rate of these two errors.
