作者: Huimin Deng , Qiang Huo
DOI: 10.1007/11527923_28
关键词: Minutiae 、 Image processing 、 Delaunay triangulation 、 Edge detection 、 Fingerprint Verification Competition 、 Artificial intelligence 、 Fingerprint 、 Computer science 、 Computer vision
摘要: This paper presents a novel minutiae matching approach to fingerprint verification. Given an input or template image, are extracted first. Using Delaunay triangulation, each is then represented as special connected graph with node being minutia point and edge connecting two minutiae. Such used define the neighborhood of that facilitates local-structure-based from fingerprints respectively. The possible alignment in can be identified efficiently. A global score between finally calculated by using aligned-edge-guided triangle procedure. effectiveness proposed confirmed benchmark test on FVC2000 FVC2002 databases.