Fingerprint image enhancement with easy to use algorithms

Abstract

This paper looks at measures to enhance image quality with the intention of improving recognition rates of finger recognition systems. There are thus different algorithms when compared to each other. The methods cover sharpness enhancement in general; wavelet sharpness, photocopy filters, EAW filters, DoG filters and Cartoon filters. The FVC2000 Db1a is used as reference dataset. The scores for the comparison are based on the NFIQ quality scores. The approach of the paper is to use a few single algorithms of the GIMP program and evaluate which one improves the reference database the most. GIMP is open-source software so everybody can use it to improve fingerprint images without having advanced programming skills.