Fingerprint matching using genetic algorithm and triangle descriptors

Because of some complex distortions which occur in fingerprint, the fingerprint matching is still known as a difficult problem. In this paper we propose a fingerprint matching method based on genetic algorithm. Since in the low quality images, minutiae extraction error is high, so in this method we have used the local properties of minutiae. Also in this paper a new kind of triangle descriptors are proposed, which uses these descriptors properties, the approximated value of transfer parameters for alignment and similarity between two fingerprints can be attained. These descriptors have been used to generate the initial population and are applied to the genetic algorithm. The results show that using this method will increase the convergence speed of genetic algorithm significantly. Also appropriate values of transfer parameters are achieved directly using this method. The proposed method in this paper has been evaluated on FVC2002-DB1_B and comparing the aforementioned results with other papers results shows that our proposed method has a better performance in convergence speed.