Prediction of Fingerprint Orientation

This paper presents an algorithm to predict fingerprint orientation. The algorithm consists of two models: coarse prediction model and polynomial model. The coarse prediction model, which is based on first order phase portrait, is proposed to estimate the orientation in areas where there is no ridge information in the input image. Polynomial model is used to refine the fingerprint orientation predicted by the coarse model. Experimental results show the effectiveness of the proposed algorithm.