Skip to Main Content
Fingerprint recognition technology is recommended to use into MRTD (Machine Readable Travel Document) by ICAO (International Civil Aviation Organization) in 2003. As most fingerprint verification/identification systems, biometric systems for MRTD are based on minutiae matching too. In this paper we divide the point pattern matching process into two stages. We propose an adjacent tri-point feature vector matching algorithm in the first matching stage. It is used to find the correspondence of the two minutiae sets to make first decision and to get more accurate parameters of alignment adjustment. The second matching stage we use a changeable matching box proposed by Luo et al. and integrate multi-discriminant to arrive the last decision. Experimental results show the performance of the proposed scheme.