Neural Network in Biometrics : A Survey in Fingerprint Classification

Fingerprint classification has become a popular research topic due to its applicability in authentication and identification. There are several approaches in fingerprint classification system which are structural, syntactic, statistical and neural networks. The purpose of fingerprint classfication is to categorize a fingerprint into certain category based on its global pattern configuration. The analysis of comparisons between neural-network and non-neural network approaches have pointed out the advantages of using neural network in fingerprint classification. The results show that the combination of neural networks with other machine learning approach outperforms the neural networks and machine learning approach is suggested in this paper. The clear advantages of supervised and unsupervised learning in neural networks methods support the objective of this study that to suggest the neural network approach for fingerprint classification. A model of neural network combined with machine learning approach (SOM-LVQ and MLP) is proposed at the end of this study. SOM-LVQ is used for pre-classification and MLP classifier is used for classification.