Fingerprint Singular-point Detection Based on Mathematical Model Analysis

Wang Xiaoduan,Liu Shuxin,Liu Manhua

DOI: https://doi.org/10.19335/j.cnki.2095-6649.2012.08.003

2013-01-01

Abstract:As traditional singular point detection methods are mainly based on the orientation changes which are often affected by noise,a novel method was proposed for fingerprint singular-point detection based on mathematical model analysis in this paper.First,discrete cosine transform(DCT) was used as the basis functions to build the orientation reconstruction model and compute the orientation field.Then,the ordinary differential equation and system linearization model were computed with fingerprint orientation field and the singular points detected based on the model analysis.Experimental results show that the proposed singular point detection method has high robustness to noise and can achieve a higher accuracy compared to the traditional Poincare Index method.