利用B样条曲线设计指纹图像滤波模板和快速指纹特征匹配算法

A Filter Creation Method Based on B-splines and a Fast Feature Matching Algorithm in Fingerprint Recognition System

在指纹图像识别过程中,指纹图像的滤波去噪音二值化和指纹特征的匹配是最关键的两个部分.本文针对这两部分的算法设计进行分析和改进,介绍利用B样条造型来设计指纹滤波模板的方法和基于基准点附近的拓扑结构进行快速特征匹配的算法.并且利用改进后的指纹特征匹配算法,在样本指纹库中进行实验,结果表明该算法大大提高了指纹特征匹配的效率和准确性.

This paper focuses on the filter creation and fingerprint feature matching algorithm, which are the two most important sections of the entire fingerprint image identification system. The first part of this paper introduces a method of creating the filter based on B-spline base functions,the second part provides a fast fingerprint feature matching algorithm based on the nearby topology structure of the minutiae points.The testing against large samples of the finger prints images shows that the algorithm effectively improves the recognition speed,and with enhanced FAR/FRR（Fail Accept Rate/Fail Recognition Rate）.