Fingerprint Classification Using Improved Directional Field and Fuzzy Wavelet Neural Network

A fingerprint classification algorithm is proposed in this paper. It is based on the features extracted from the directional field of the fingerprint image. To improve the accuracy of the directional field, an efficient estimation approach is developed. Based on the improved directional field, the singular points and the relative features are extracted to generate the input features of the fingerprint classifier. After encoding the input features, a fuzzy wavelet neural network-based classifier is applied to classify fingerprints for the five-class problem. Experimental results show an excellent classification performance of the proposed algorithm