作者: Jiang Li , Sergey Tulyakov , Zhi Zhang , Venu Govindaraju
DOI: 10.1109/BTAS.2008.4699319
关键词:
摘要: One of the difficulties for fingerprint matching lies in non-linear distortion between images originating from same finger. In this paper we present a modification correlation method, which uses thin-plate spline (TPS) as model transformations two fingerprints. The TPS is constructed first results traditional minutia algorithm, and by using one fingerprints deformed to match other. After transformation scores local neighborhood areas corresponding minutiae pairs edges that connect neighboring matched are calculated used final score. FVC2002 DB1 database test proposed approach. Experimental result shows improvement when combining deformation with method.