NIST releases latent fingerprint biometrics training data, quality assessment software

New biometric training data and open source fingerprint quality assessment software has been released by NIST to help improve the performance of matching software and print examiners alike.

NIST Technical Note (TN) 2367 describes Special Database (SD) 302 as a resource of “Annotated Latent Distal Phalanxes.”

The paper was composed by a team from the Information Access Division within NIST’s Information Technology Laboratory and another from Schwarz Forensic Enterprises.

The SD 302 fingerprint biometric dataset consists of 10,000 images gathered from 200 volunteers. NIST computer scientist Greg Fiumara explains in the announcement that the data was acquired by having the volunteers handle a set of everyday items, and then scientists collect the prints they left behind through methods commonly used by crime scene investigators.

SD 302 was originally collected as part of NIST and Intelligence Advanced Research Projects Activity’s (IARPA’s) Nail to Nail Fingerprint Challenge.

The dataset was released in December, 2019, and has been updated several times since, according to the paper. Annotating the data has taken years, however, and a version released in November of 2021 only included annotations for about half of the images. The remainder have been annotated in the latest release. Additionally, SD 302 is broken down into nine datasets, each with different print types or characteristics. These datasets are referred to as SD 302a-i.

“These images are good for classroom education, to teach examiners how to look for identifying features,” says Fiumara, one of the report’s authors. “And they will also help teach AI algorithms where to look and how to weigh a feature’s importance. With this kind of training, a fingerprint evaluation algorithm will get better.”

Introducing OpenLQM

LQMetric was developed from 2012 to 2014 by American federal government research body Noblis, with funding from the FBI Criminal Justice Information Services (CJIS) Division and the Bureau’s Center of Excellence.

“The value returned by LQMetric is an estimate of the probability that an image-only search of the Federal Bureau of Investigation’s (FBI) Next Generation Identification (NGI) automated fingerprint identification system (AFIS) would hit at rank 1 if the subject’s exemplar (rolled) fingerprints are enrolled in the gallery,” explains a 2020 research paper from Noblis.

The LQMetric fingerprint evaluation software was previously limited to use by U.S. law enforcement, but NIST funded a version that will run on Mac, Linux and Windows operating systems over the last year, according to the announcement.

The result is OpenLQM, which can run as a standalone program or be integrated with other software as a plug-in.

“You give OpenLQM a fingerprint and it returns a number from 0-100 that is an assessment of the print’s quality,” Fiumara says. “It can help print assessors work more quickly, which is important in forensic science when you often have hundreds of prints to review from a crime scene. You want to help them separate out the prints that contain the highest level of detail.”

Anthony Koertner, a certified latent print examiner at the Department of the Army Criminal Investigation Division’s U.S. Army Criminal Investigation Laboratory, says the open-source release and SD 302 together are “a significant advancement for the global forensic community” that has helped his Department’s “efforts to achieve greater objectivity and reproducibility in latent print quality assessments.”

Australia’s BixeLab launched its software tool for biometric data quality investigation, BQAT, as open-source software OpenBQ in February at MOSIP Connect 2026.

Article Topics

biometric data quality  |  biometric dataset  |  biometric matching  |  biometrics  |  fingerprint biometrics  |  NIST  |  OpenLQM