Fingerprints are an important biometric trait, being the most widely deployed biometric character... more Fingerprints are an important biometric trait, being the most widely deployed biometric characteristic. Minutiae-based methods are the most popular technique to compare fingerprints, but as the pairing is rarely perfect the score between two fingerprints may also depend on another factor. Usually, a constant weight is computed either empirically or statically and assigned for each factor. Optimize these weights can be a hard task for a large N-dimensional attributes space. Genetic Algorithms (GA) is an optimization approach based on Darwin's theory of evolution of species and it has been proved to be quite successful in finding good solutions to such complex problems. In this work, we analyze an GA-based optimization approach for SourceAFIS's fingerprint authentication algorithm. The DB2 fingerprint database from FVC 2006 project was used to validate our method. As best result, an EER of 0.379% was achieved.
Fingerprints are an important biometric trait, being the most widely deployed biometric character... more Fingerprints are an important biometric trait, being the most widely deployed biometric characteristic. Minutiae-based methods are the most popular technique to compare fingerprints, but as the pairing is rarely perfect the score between two fingerprints may also depend on another factor. Usually, a constant weight is computed either empirically or statically and assigned for each factor. Optimize these weights can be a hard task for a large N-dimensional attributes space. Genetic Algorithms (GA) is an optimization approach based on Darwin's theory of evolution of species and it has been proved to be quite successful in finding good solutions to such complex problems. In this work, we analyze an GA-based optimization approach for SourceAFIS's fingerprint authentication algorithm. The DB2 fingerprint database from FVC 2006 project was used to validate our method. As best result, an EER of 0.379% was achieved.
Uploads
Papers by Pedro Alves