LQMetric: A Latent Fingerprint Quality Metric for Predicting AFIS Performance and Assessing the Value of Latent Fingerprints

We describe LQMetric, an automated tool for measuring the image quality of latent fingerprints. The value returned by LQMetric is an estimate of the probability that an imageonly search of the Federal Bureau of Investigation's (FBI) Next Generation Identification (NGI) automated fingerprint ide...

Full description

Saved in:
Bibliographic Details
Published inJournal of forensic identification Vol. 70; no. 4; pp. 443 - 463
Main Authors Kalka, Nathan D, Beachler, Michael, Hicklin, R Austin
Format Journal Article
LanguageEnglish
Published Alameda International Association for Identification 01.10.2020
Subjects
Automation
Criminal investigations
Forensic sciences
Latent fingerprints
Online AccessGet full text

Cover

More Information
Summary:We describe LQMetric, an automated tool for measuring the image quality of latent fingerprints. The value returned by LQMetric is an estimate of the probability that an imageonly 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. LQMetric can also be used in assessing the value of latent fingerprints in non-AFIS casework. LQMetric is incorporated into the FBI's Universal Latent Workstation (ULW) software and has been used operationally since 2014. The development of an automated latent fingerprint quality metric was driven by practical use cases including predicting the likelihood of successful AFIS matching; helping examiners determine whether an image-only or human-markup search is more appropriate for a particular latent fingerprint; supporting a quality-directed workflow whereby a backlog is prioritized based on quality or lower quality latent prints are directed to highly experienced examiners; or providing an objective difficulty measure for quality assurance purposes such as flagging complex prints for special handling or additional verification. We describe how LQMetric was developed and trained, how well it predicts NGI AFIS search results, and we also discuss human examiner latent fingerprint value assessments.
ISSN:0895-173X