Fingerprint Classification in DCT Domain using RBF Neural Networks

Fingerprint classification is a fundamental method for the identification of people. Fingerprint classification is based on the immutability and the individuality of fingerprint. Because of the large collections of fingerprints and recent advances in computer technology, there has been increasing interest in automatic classification of fingerprint. In this paper, an efficient method for fingerprint classification based on the discrete cosine transform (DCT), fuzzy c-means clustering (FCM), the Fisher’s linear discriminant (FLD) and radial basis function (RBF) neural networks is proposed. Experimental results show that the proposed method achieves excellent performance with high correctly recognition rate, very low reject rate, and very less running time.