基于梯度矢量的指纹特征检出算法研究  

A Study of Gradient based Fingerprint Feature Detection

作　　者：何小海[1] 罗代升[1] 滕奇志[1] 陶青川[1] 冉均国[2] 

机构地区：[1]四川大学电子信息学院,成都610064 [2]四川大学材料科学与工程学院,成都610064

出　　处：《计算机工程与应用》2003年第2期20-23,共4页Computer Engineering and Applications

基　　金：四川省自然科学基金资助

摘　　要：自动指纹识别的关键技术之一是指纹特征的检出。目前,指纹特征检出大多是先采用图像增强和图像分割技术检出指纹纹线,然后从指纹纹线图像中检出指纹特征。用这些方法检出的指纹特征依赖于图像分割的精度,而且检出过程比较复杂,运算速度比较慢。为此,提出一种直接按指纹梯度矢量的方向图进行指纹特征检出的算法。文中的算法是先采用拓宽的Prewitt算子计算指纹图像的梯度矢量图,然后计算梯度矢量的方向一致率,最后由方向一致率图检出指纹特征点。首先介绍方法的基本原理,然后给出了采用提出的方法检出指纹特征的实验例子。One of the key techniques of automatic fingerprint i de ntification is fingerprint feature detection.Today,most of the fingerprint f eature detection is first to detect fingerprints by image enhancement and segme ntation,then to detect the fingerprint features from the fingerprints image.T he features detected by these methods depend on the accuracy of image segmentati on.The procedures of the detection are complex and the speed is slow.For this reason,the authors propose a method to detect fingerprint features directly fr om the fingerprint gradient orientations.In their method,first,a widened Pr ewitt operator is applied to calculate the gradient vectors of a fingerprint i mage.Then,the consistency ratio of the gra-dient orienta tions is calculated .Finally,the fingerprint features are detected from the consistency ratio ima ge.In this pa-per,the principle of our method is first introduced.Then an e xample of the experiments of fingerprint feature detection by the method is giv en.

关 键 词：梯度矢量 指纹特征检出算法 指纹识别 指纹方向图 图像增强 图像分割 计算机 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] D918.91[自动化与计算机技术—计算机科学与技术]