Secure Vascular Biometric Recognition
Chris Humphry, Sunil Rufus Ramneedee Pushparaj, Nalini Ratha · 2023
The use of biometric technologies for personal identification and authentication has become increasingly popular in recent years. Among these technologies, vascular feature recognition is considered one of the most secure and reliable methods due to its ability to capture the unique biological characteristics of the veins beneath the skin’s surface. In this paper, we propose an end-to-end vascular feature recognition system built with the security and privacy of the template in mind. Our solution uses Resnet50 architecture coupled with biohashing to securely store and manage the biometric data of individuals. The system makes use of the images captured using a near-infrared light scanner. This image is processed to generate a biohash, a unique digital representation of the user’s vein pattern. The system also employs a matching algorithm that compares the user’s biohash with the stored biohash to verify the user’s identity. The core model achieved an accuracy of 95.58% compared to the state-of-the-art accuracy of 97.66%. However, our proposed solution achieves an EER rate of 0 after biohashing whereas the state-of-the-art model has an EER rate of 2.03, indicating 100% efficiency of biohashing.