TECH5 has received a US patent for the neural network-based biometric scanning and liveness technology used in its T5-AirSnap Finger solution. Patent 11721120, titled ““Contactless fingerprint capture using artificial intelligence and image processing on integrated camera systems”, details a system that, among other things, produces a 3D depth map of a finger enabling 500 dpi resolution.
Neural network technology plays an important role in the system described. NN models are first trained to detect the size and direction of fingers captured in a given image and whether the nails are visible, among other metrics; and then to generate fingerprint templates of a fixed length that can be used for subsequent matching. Generating a 3D depth map helps to eliminate any distortion caused by a fingerprint’s curvature, which in turn improves accuracy.
The entire process is designed to happen on a standard smartphone, eliminating the need for the use of specialized hardware in contactlessly capturing and scanning fingerprints.
“Our expert R&D team is committed to advancing contactless fingerprint capture and liveness detection, integral to critical applications such as national enrollments, law enforcement, eKYC, and decentralized digital ID,” said TECH5 CTO, Chairman, and co-founder Rahul Parthe. “We take pride in leading the way in this exciting new field of research.”
TECH5 first unveiled its T5-AirSnap Finger solution in 2021, launching in in the form of a software development kit. The solution had arrived at a time of heightened concern about contact-based fingerprint scanning, due to the ongoing pandemic; but TECH5’s biometric technology has continued to see interest, with one high-profile deployment currently underway in an ambitious digital ID project in Ethiopia.
News of the patent approval for T5-AirSnap’s underlying technology comes after TECH5 attained top rankings in a contactless fingerprint evaluation program organized by Clarkson University and the University of Buffalo.
–
(Originally published on FindBiometrics)
Follow Us