广东工业大学学报 ›› 2005, Vol. 22 ›› Issue (3): 43-47.

• 综合研究 • 上一篇    下一篇

一种提取指纹模式区几何框架的新算法

  

  1. 广东工业大学计算机学院; 广东工业大学计算机学院 广东广州510090; 广东广州510090; 广东广州510090;
  • 出版日期:2005-07-02 发布日期:2005-07-02
  • 基金资助:

    广东工业大学博士启动基金项目(053017)

An Effective Algorithm for Extraction of Fingerprint Geometric Framework

  1. (Faculty of Computer,Guangdong University of Technology,Guangzhou 510090,China)
  • Online:2005-07-02 Published:2005-07-02

摘要: 提出了一种稳健的提取指纹模式区总体几何形状的有效算法.算法中采用了指纹方向场估计的巧妙处理方法,即使在局部纹线方向发生较大变化的情况下,这种方法也能使跟踪提取的总体几何框架保持基本稳定.在跟踪处理过程中,采用了一种自适应跟踪方法,而且只在被跟踪点处估计纹线的点方向,因而算法的操作时间相对较少.该算法在NJU指纹数据库中的500枚指纹上进行了验证,实验结果表明算法是有效的. 

关键词: 指纹; 总体几何形状; 伪纹线; 方向场估计; 自适应跟踪;

Abstract: In this paper,a robust pseudoridges extraction algorithm for fingerprints is presented to gain the global geometric shape of fingerprint ridges of pattern area.The algorithm adopts the skillful processing method for orientation field estimates,so that the pseudoridge traced remains constant under large variations of local ridge orientation.Hence it is more robust than the feature of the specific ridges.In the tracing process,a method with adaptive tracing and estimating the orientation only on the traced point is presented to reduce the operation time.The algorithm for pseudoridges extraction has been tested on the 500 fingerprints of NJU fingerprint database,and gets the good performance according to the experimental results.

Key words: fingerprint; global geometric shape; pseudoridge; orientation estimate; adaptive tracing;

[1] 尹义龙,宁新宝,张晓梅.  自动指纹识别技术的发展与应用[J]. 南京大学学报(自然科学版). 2002(01)

[1] Rao K,Black K.Type classification of fingerprints: A syntactic approach. IEEE Transactions on Pattern Analysis and Machine Intelligence . 1980

[2] KASS M,WITKIN A.Analyzing oriented patterns. Computer Vision . 1987

[3] Chong M M S,Ngee T H,Jun L,Gay R K L.Geometric framework for fingerprint classification. Pattern Recognition . 1997

[4] Kawagoe M,Tojo A.Fingerprint pattern classification. Pattern Recognition . 1984
No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!