改进的指纹奇异点检测算法  

Improved Algorithm for Fingerprint Singular Points Detection

作　　者：公绪成[1] 孙立民[1] 郑植庆[2] 

机构地区：[1]烟台大学计算机科学与技术学院,山东烟台264005 [2]山东工商学院信息网络中心,山东烟台264005

出　　处：《烟台大学学报（自然科学与工程版）》2010年第2期132-137,共6页Journal of Yantai University(Natural Science and Engineering Edition)

基　　金：山东省自然科学基金资助项目(Y2006G22);山东省信息产业发展专项基金资助项目(2007D00061)

摘　　要：为解决传统指纹奇异点检测算法不能检测拱型、计算量大等缺点,提出了一种拱型指纹奇异点定义和一种新型的类Poincare指数,并给出了基于类Poincare指数的改进的奇异点检测算法.首先计算指纹奇异点可能出现的区域,然后根据可能区域内类Poincare指数值的变化曲线精确确定奇异点的位置.引入了可能区域,降低了检测过程中的计算量,因此大大提高了提取奇异点的速度.该方法也适用于其他类型的指纹.在FVC2004指纹库上进行实验,能准确检测拱型和其他类型指纹的奇异点,且检测速度大大提高,证明该方法可行有效.Traditional method for singular points detection can not detect an arch fingerprint, and requires a large amount of calculation. In order to conquer the shortcomings, a new definition for arch fingerprint singular point is presented. A new index named Poincare-like index and an improved detecting algorithm which is on Poincare-like index are brought forward. Potential area where singular points are most probably to appear is firstly computed, the accurate positions of singular index curve. Calculation amount is reduced because process is greatly sults on FVC2004 speeded points are determined according to the Poincare-like of the introduction of potential area, therefore, detecting up. This method can also be applied on fingerprints of other types. Experiments redatabases show that this method can accurately detect singular points on fingerprints of arch and other types, and detection process is highly speeded up.

关 键 词：指纹 奇异点检测 类Poincare指数 可能区域 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]