摘要
随着指纹识别技术的广泛应用和网络技术的发展,基于网络的指纹自动识别系统面临着如何提高检索速度和准确度的问题。本文根据指纹数据库自身的特点,提出了一种新的检索和索引网络大型指纹数据库的方法。该算法利用指纹的类别和局部特征建立了三级索引,有利于缩小检索空间,提高检索速度。在算法中提出了一种新的指纹分类方法,将指纹分为八类:弧形、尖拱形、左旋形、右旋形、正涡形、左涡形、右涡形、混合形。该算法利用core点和delta点的位置、数目和方向判断指纹的类别,有效地解决了采集时对指纹的旋转、平移和形变不变性的要求,并有助于进一步缩小检索空间,提高检索速度。实验证明,效果良好。
With the wide use of the fingerprint recognition and Internet development,FAIS （fingerprint automatic identification system）based on Internet confronts how to improve search rate and accuracy. On the basis of the features of the fingerprint database, this paper proposes a new method of searching and indexing the large net fingerprint database. This method uses the fingerprint classification and regional features to set up the three-level index. It helps reduce the search space and increase the search rate. Meantime, a new fingerprint classification way is presented in this paper. Fingerprint is classified into eight categories：arch, tented arch, left loop, right loop, vertical whorl, left whorl, right whorl, mixture. This classification algorithm not only uses the location and number of core points and delta points, but also uses their direction. It successfully solves the problem of fingerprint rotation, translation and transformation, and further reduces the search space and improves search rate. The experiment proves that the search way is efficient.
出处
《计算机科学》
CSCD
北大核心
2005年第8期196-199,共4页
Computer Science
