Spectral selection for a biometric recognition system based on hand veins detection through image spectrometry

Abstract

This paper presents the result of a work orientated to the spectral optimization of the acquisition devices in vascular biometrics systems. Spectral windows are proposed which will allow to design a multispectral system with a few and well defined bands, obtaining a more robust and reliable device, compared with the standard single band systems. This is in accordance to general trend of electro-optical and infrared acquisition systems in the field of the detection and remote sensing, where the work focus is on obtaining optimized bands. To carry out this work a Hyperespectral Imaging System (HIS) has been used as the acquisition system. In order to analyze the large amount of information and to select the spectral bands, a Principal Component Analysis (PCA) has been done.