Whole-hands multiple-instances finger vein biometric system

Abstract

This paper introduces a discipline to improve the finger vein biometric system accuracy and performance by incorporating multiple finger vein patterns for each person. Instead of just a single finger, we acquired 2 fingers from the right hand and 2 from the left hand to represent a single identity. Firstly, each finger region was segmented then the vein tree for each finger was extracted using maximum curvature points in image profiles. The binary vein pattern for each single finger was matched using the phase only correlation (POC) technique. The score level fusion methodologies were used to obtain a single decision from multiple finger vein patterns belonging to a single identity. The designed system can be considered for authentication and identification purposes. Advantages of this system over few existing multimodal systems are its being very hard to spoof attacks on the sensory level and the NIR FV thermal images are good signals for liveness detection and ensuring.