An Energy-Based Fingerprint Matching System

Khaled Ahmed Nagaty
Faculty of Computers and Informations Sciences
Ain-Shams University, Cairo, Egypt

Abstract:
In this paper we present a new energy-based fingerprint matching system which uses both minutiae information available in a fingerprint with the information of the local ridges in their vicinity. The basic idea of this system is to divide the fingerprint matching problem into several small sub-problems that involve the use of image energy minimization for which an iterative schema is devised. At each minimization step this schema optimizes its local energy according to the previous estimate and the observed image features. Different amounts of energy are produced by using minutiae and the local ridges in their vicinity which form the energy residuals of the fingerprint image. In this work, we focus on the difficult problem of recognizing known fingerprints while rejecting unknown ones. Our system was tested on NIST database of fingerprints and showed promising results. We show that matching performance can be improved by using energy vectors which can efficiently represent fingerprints by combining both the information extracted from the minutiae and their local surrounding ridges.
Keywords: Fingerprints, Minutiae Matching, Minutiae Weights, Directional Weights, Energy Minimization, Nonlinear Deformations