Design and Implementation of Vascular Pattern Recognition System

Abstract

This paper proposes a novel technique to analyze the vein patterns in the back of the hand for biometric authentication system and medical purposes. It deals with the design development of non-invasive subcutaneous vein detection system and is implemented based on near infrared imaging and interfaced to a PC to make it portable. A web camera is used for capturing the images and Computer Software modules (MATLAB & SIMULINK) is used for the processing. The first task is to extract the hand and then the vein patterns. To identify the patterns in hand, near-infrared rays generated from a bank of light emitting diodes (IR LEDs) penetrate the skin of the Palma dorsa. An image of hand veins acquired by near-infrared (NIR) imaging device, usually suffers from low contrast and noise due to non-illuminated and thickness of the hand skin. This makes subsequent processing such as segmentation difficult. To enhance the quality of hand vein image we used the threshold segmentation method. Extensive experiments were carried out on dorsal veins to achieve the promising accuracy and thereby provide new insights on this new biometric approach.