Fingerprint image enhancement based on elliptical shape Gabor filter

Abstract

A key step of the AFIS (Automatic Fingerprint Identification System) is the fingerprint image enhancement. In this paper, an algorithm based on elliptical Gabor filter is presented. The method can determine the range of filtering dynamically by collecting the ridge information on the fingerprint. This is accomplished via estimating the degree of curvature and the frequency of fingerprint ridge in local areas so as to correct and more precise enhancement while avoiding the block effect. The experimental result shows that the precision of minutiae extraction are significantly improved, which directly resulted in a higher accuracy rate of the subsequent operation for fingerprint matching.