Four Clarkson students competed in the national Nail to Nail (N2N) challenge to improve live and forensic biometric fingerprint recognition. They made it to Phase 3 of the Intelligence Advanced Research Projects Activity (IARPA) Nail-to-Nail Fingerprint Challenge in Washington, D.C.
Currently, human operators must physically “roll” the subject's fingerprints over a surface in order to capture the entire fingerprint from one edge of the nail bed to another. The objective is to produce an automated capture technology that can get rid of a human operator physically interacting with the subject, and the automated technology must be as good as, or better than the traditional rolled capture. A human must be present, but there is to be no physical interaction.
The team members include Jacob Cary ’18, Software Engineering, Benjamin Lowit ’18, Computer Engineering, and Andrew Davis ’19, Computer Engineering. The team leader was David Yambay, Ph.D. student, Electrical Engineering. Faculty Advisors were Sean Banerjee, Natasha Banerjee, and Stephanie Schuckers.
The work is sponsored by the Center for Identification Technology Research (CITeR), an NSF center.
The students have prototyped a non-contact fingerprint scanner from 15 raspberry pi computers and cameras, and are utilizing state-of-the-art 3D reconstruction to create a 3D representation of the finger which can be used converted to a 2D nail-to-nail fingerprint.
The Clarkson team attended the competition with the hopes of collecting data. Work on the project began at the start of Clarkson’s summer session. Over the summer, the students worked countless hours to create a system for the competition. The system was a prototype that was designed for collecting data and producing simple images. The output images of the system weren’t quite ready for fingerprint matching but taught the students a lot so that they could continue the project through May 2018.
The companies present at the competition were impressed by the system and happy to see Clarkson participating and showing off the work that the undergraduate students could accomplish. The test itself was grueling. The group collected images from over 350 individuals over the course of five days on a strict timeline of five minutes per subject throughout the day. Work on the project will continue throughout the school year, and a fully functioning system is expected to be completed in the spring semester.