Versatile and Economical Acquisition Setup for Dorsa Palm Vein Authentication

摘要

Various biometrics were employed in many applications for security purposes, amongst palm vein biometrics is one of the best methods for unique identification of a person owing to the indestructible quality of the inner vein structures. In this paper, we have proposed our own setup for capturing vein structures of human dorsal palm using a web camera modified into a near infrared camera. The illumination for capturing images is provided with the help of 30 Infrared LEDs. The objective of this paper is to produce a versatile and an economical way for obtaining vein images rather than using a high priced Near Infrared Camera and can easily deployed in any small scale applications. This setup can be used to acquire finger veins too simultaneously. We have modified the web camera by removing the infrared filter present in it and replacing it with a visible light filter. The quality and performance of the newly acquired samples are tested with two different feature extraction methods namely Correlation filter and Speeded Up Robust Features (SURF) algorithm. Correlation method has obtained very good results than SURF in identifying the genuine samples.