NIST report notes improved rolled fingerprint systems
09 May 2018 17:02 GMT

Fingerprint capture technology has advanced to the point where high-quality rolled prints soon might be obtained by machines alone, according to a new report by the National Institute of Standards and Technology (NIST).

The document, NIST Interagency Report (NISTIR) 8210, Nail to Nail Fingerprint Challenge: Prize Analysis, details the methods used in a recent IARPA-sponsored challenge whose overall goal was to improve fingerprint capture technology. Eight finalists from industry and academia took part in the competition’s finale, the results of which IARPA announced at a March 21 workshop.

NIST’s role was to design the experiment and evaluation criteria so that IARPA could determine the prizewinners. The report details this background for organizations that may want to work on their own approaches to the problem, which is not yet fully solved despite the advances the challenge delivered.

“One of the competitors brought a solution that can provide images at the capture speed, reliability and typical quality level of the traditional, operator-assisted system,” said NIST’s Elham Tabassi, one of the report’s co-authors. “The one caveat is that someone was watching and providing feedback to the user from a distance.”

In this video created by IARPA, participants and organizers of the Nail2Nail Challenge discuss the effort to develop technology that can obtain high-quality fingerprints from a human subject without requiring a trained operator’s manual assistance.

The ideal solution would require no assistance whatsoever, but even limiting the requirement to verbal guidance would represent an improvement over conventional print-capturing methods. Obtaining a high-quality image of an entire fingerprint—from one side of the nail around the fingertip to the nail’s other edge, or “nail to nail”—is a demanding task. It typically requires the assistance of a skilled operator who physically grasps and rolls the person’s fingers, a time-consuming process that can make the person feel awkward.

“Requiring a skilled operator in a high-speed environment like a port of entry is costly, and it can constrain print collection,” Tabassi said. “Plus, for many civilian applications, plain finger impressions provide sufficient information for verification of identities, so many real-world applications end up taking plain fingerprint impressions instead.”

Plain impressions are the commonplace sort that only need a flat press of the finger, like on a modern smartphone. While they are easier to obtain without assistance, they lack detail from the edges of the finger that can be important when searching through a database of prints taken from crime scenes, for example. These unintentional (or latent) prints are often partial or off-center, showing the sides of a fingertip or bits that only a nail to nail scan would capture.

“The challenge’s overall results show that the winning device is ready to produce images,” Tabassi said. “The remaining issue is how to ensure that the device can produce high-quality prints without the verbal feedback from a moderator.”