Innovatrics and Idemia top latent fingerprint biometric accuracy ranks

Innovatrics is claiming a win in the Evaluation of Latent Fingerprint Technologies (ELFT) program from the National Institute of Standards and Technology, with biometric accuracy tied for the highest score in rank-1 hit rate with Idemia, and alone in first for rank-5 reviews.

Since June 1, 2024, new submissions have come from NEC, ROC, Innovatrics, Hisign Technology, and Idemia. By false negative identification rate (FNIR) at a false positive identification rate (FPIR) of 0.01 on an FBI dataset, Idemia was most accurate, at 0.0795.

Innovatrics achieved a 96.5 percent hit rate in rank-1 searches, and a 97.5 percent hit rate (0.0388 FNIR) at rank-5 in the solved dataset #1 provided by the FBI. Additionally, the company has been recognized as the most accurate in the rank-based search over the newly added Department of Defense dataset #1.

The second closest algorithm in the rank-5 in the FBI solved dataset #1 is from Idemia. NEC and ROC have scores of 0.0775 and 0.1182 FNIR, and 95.7 percent and 91.5 percent hit rates, respectively.

Innovatrics also says that it has improved lights-out searches, where a threshold is applied to filter out non-matching candidates. The algorithm demonstrates a rank-100 hit rate of over 91 percent at a false positive identification rate (FPIR) of 1 percent.

In automated fingerprint identification systems, lights-out refers to automated searches in which the computer compares latent fingerprints against a large database without human intervention during the initial matching process.

Idemia had scored the best accuracy in rank-1 hit rate among ELFT submissions as of March, with the top speed among the competition’s most accurate algorithms.

Article Topics

accuracy  |  biometric testing  |  biometrics  |  ELFT  |  fingerprint biometrics  |  IDEMIA  |  Innovatrics  |  NIST