Automated Fingerprint Matching Algorithm using Representative Ridge Points

Abstract

Automatic fingerprint identification systems utilize algorithms which matches only extracted minutiae from the input fingerprint- which are ridge endings and ridge bifurcations with minutiae in known fingerprints. Verification error rates of such algorithms occurs especially if the number of available minutiae are few which are not enough to give a reliable matching. Existing algorithms use local ridge similarity only for fingerprint alignment which are sensitive to non-linear deformation (since ridges cover relatively larger regions). An algorithm is proposed to utilize ridge information that performs more effectively by choosing representative points along the ridges. The chosen ridge points are used, together with minutiae, in existing minutiae matching algorithms. Since the addition of ridge points increases computation time significantly, a ridge point selection scheme is proposed, through which only one representative ridge point (RRP) is selected per ridge. The approach has two merits: significant performance improvement for fingerprints and usability with conventional matchers.