Infrared hand dorsum images for identification

Abstract

A quick and simple personal identification scheme based on the infrared image of the hand dorsum is presented. The process produces a template that is obtained by filtering an extended infrared hand dorsum image with a Gabor filter. Therefore, it includes vascular and geometrical information of the hand. A matcher based on Hamming distance decides to accept or reject the user. The scheme was tested on an operational database of 150 users and resulted in an averaged equal error rate of 1.43%.