Fingerprint Classification Based on Spectral Features

Fingerprint is one of the most important indexes that can be applied for verification and identification. In the recent decade, with the development of societies and databases of fingerprint, automation of identification has been unavoidable. Fingerprint classification decreases the time of search for an unknown image in large databases. The purpose of this research is to increase the number of classes and improvement the accuracy of classification. Fingerprint images are classified into seven classes: Right loop, left loop, Twin loop, Arch, Tented arch, Whorl and Central packet loop. In this research, translation invariant features are extracted from spectrum of the fingerprint image. The extracted features obtain not only information from frequency of ridges but also valuable information from direction of ridges in the fingerprint images. Features are classified with Probabilistic Neural Network. FVC2000 and FVC2002 databases are used to assess the proposed algorithm. The proposed algorithm provides an accuracy and speed of classification better than previously reported in the literature.