Fingerprint verification system: A non-minutiae based approach

In this paper, a new method for fingerprint verification, based on the features extracted from both spectrum and direction domains of fingerprints, is introduced. In this method, the spectrum of fingerprint is divided to sectors with equal number of pixels and for each sector, the respected average and standard deviation is calculated. The resulting feature vector for any fingerprint is used to calculate the distance measure of two fingerprints. Also a value is assigned to the similarity of the Block Directional Field (BDF) of two fingerprints. The fusion of the above mentioned distance and similarity of BDF's is used to verify an input fingerprint image. This method does not need any preprocessing step and due to simplicity, the speed of verification is high. Also, in spite of other image-based methods, this method does not need any reference point to be detected. The experiments show better results for verification accuracy in comparison with the most important methods presented in the literature.