An Effective Fingerprint Classification using Genetic Algorithm and Markov Models

The existing fingerprint classification methods are well suited for fingerprints acquired using paper and ink. However, those are not so efficient with recent automatic fingerprint systems because it cannot guarantee that singular points are well extracted since the recent systems have various sized sensors and use multifarious fingerprint acquisition methods. In this paper, a novel approach is proposed to use the fingerprint ridge directions, which are one of the global features. We make the model of each class using the extracted directional characteristic of fingerprint ridges, fingerprint models were built with the Markov model, and the generated Markov model of each class was optimized by genetic algorithm. After that, Markov model of each class was modified. We use FVC2000 DB1 and FVC2002 DB1 database to classify fingerprints and analyze the classification result. This approach can be effectively applied when the quality of fingerprints is poor or a partial image is classified.