Chapter 20: A Fingerprint Local-Matching Algorithm Using Unit-Circle Parametrization

Abstract

Pattern recognition provides solution to many problems in real life such as in biometric system, personal identification of banks etc. It matches two point sets and consequently identify if they are identical. This is applicable in fingerprint recognition with minutiae as a representation, which has been widely used as an individual identification method. Fingerprint recognition is divided into two parts. One is to extract feature points from fingerprint image, another is matching of point pattern. This paper presents a matching algorithm. The Wamelenpsilas approach, which finds k-nearest neighbors, is quite famous recently. But in this paper, we studied the application of Delaunay triangulation and parametrization. This method maps local neighborhood of points of two different point sets to a unit-circle. In this paper, we get topology information, which is the raw data, from feature point of real finger by using Delaunay triangulation method. In consisted topology structure, we call a linked convex polygon that includes an interior point as one-ring. The one-ring is mapped to a unit-circle using parametrization. In this paper, we use shape-preserving parametrization method. In local matching, each area of polygon in unit-circle is compared. If the difference of two areas are within tolerance, two polygons are consider to be matched and then translation, rotation and scaling factors for global matching are calculated.