指纹曲线数值拟合方法的对比和优选

Comparison and optimization of numerical fitting methods for fingerprint curves

指纹识别在数据加密方面有广泛的应用,在此基于数值方法,借助Matlab软件,利用最小二乘拟合、多项式曲线拟合、单项式基本插值、拉格朗日基本插值、牛顿基本插值、分段多项式插值和样条插值等不同曲线拟合方法对指纹曲线分别进行拟合,并对拟合效果和算法进行比较,优选最佳方法。针对劣性情况的指纹曲线,将广义延拓逼近法应用于指纹曲线拟合中,进一步完善指纹曲线拟合,实现了各种拟合方法的总结和比较以及扩展。

Fingerprint identification,as a popular identification method,has been widely used in data encryption. The different curve fitting methods such as the least-squares fitting,polynomial curve fitting,monomial interpolation,Lagrange interpolation,Newton interpolation,polynomial interpolation in subsection,spline interpolation and so on are utilized in this paper to make the fingerprint curve fitting by means of Matlab software and on the basis of numerical methods. And then the fitting results and algorithm are compared to select an optimal method. The generalized extension approximation method is used in fingerprint curve fitting to deal with the bad fingerprint curves to further improve the fingerprint curve fitting. The summary,comparison and expansion of various fingerprint curve fitting methods were achieved.