Finger Vein and Knuckle Print Identification System

Clients

Techstack

Purpose

Passwords and ID cards can be lost, stolen, or shared, but the patterns inside your own fingers cannot. This system set out to identify people from their finger vein and knuckle patterns, combining two hard-to-forge biometric traits for recognition more reliable than either one alone.

Description

We built a capture rig that photographs the finger under infrared light and sharpens the vein and knuckle detail, then feeds it to two purpose-trained deep-learning models, one specialized for veins and one for knuckle prints, chosen by benchmarking several candidates to find the best-performing pair. A desktop interface handles enrollment and live identification, driving the camera and illumination itself. Building a dual-trait biometric system accurate enough to trust for identity is exactly the kind of applied computer-vision R&D we take on.