Skip to content
Finger Vein and Knuckle Print Identification System — AI & Machine Learning project by YenkoDev

Finger Vein and Knuckle Print Identification System

Software as a Service

No

Category

AI & Machine Learning

Clients

Confidential

Techstack

TensorFlowKerasPythonOpenCVPyQt5Raspberry Pi

Purpose

Passwords and ID cards can be lost or shared; vascular and knuckle patterns cannot. This system identifies individuals from finger vein and knuckle print images, pairing two biometric traits for more reliable recognition than either alone.

Description

An infrared-illuminated capture rig on a Raspberry Pi photographs the finger, and CLAHE preprocessing sharpens vein and knuckle detail before classification. Two convolutional networks — ResNet50V2 for finger veins and MobileNetV2 for knuckle prints — were trained in TensorFlow and benchmarked against VGG19 alternatives to select the best-performing pair. A PyQt5 interface handles enrollment, capturing 25 samples per user, and live identification, with the application driving the white and IR LEDs over GPIO.