一种通过方向延伸向量检测指纹细节点的方法  

Detection of fingerprint minutiae using direction extension vector

作　　者：佟喜峰[1] 刘松波[1] 黄剑华[1] 唐降龙[1] 孙宁[2] 

机构地区：[1]哈尔滨工业大学计算机学院,哈尔滨150001 [2]大庆石油学院,黑龙江大庆163318

出　　处：《哈尔滨工业大学学报》2008年第5期750-753,共4页Journal of Harbin Institute of Technology

基　　金：国家自然科学基金重点资助项目(60332010)

摘　　要：为了提高基于细节点的指纹鉴别系统的性能,提出了将方向延伸向量特征用于细节点检测,定义方向延伸向量为在预先定义的各个方向上在相同亮度条件下最大延伸长度所组成的向量.细节点检测包括4个步骤:通过方向延伸向量检测出所有可能的细节像素点;通过计算局部区域内的细节像素点的平均位置来获取细节点位置;通过搜索方向延伸向量中最大值对应的方向获取细节点方向;根据细节点间的相对位置和相对方向消除伪细节点.实验结果表明该方法能够快速而准确地检测细节点.To improve the performance of a minutiae-based fingerprint verification system, we propose a feature called direction extension vector （DEV） for fingerprint minutiae detection, each item of which can be described as maximum extension distance from a pixel in each pre-defined direction in the sub-region with the same intensity as the pixel. The fingerprint minutiae detection approach consists of four procedures. Firstly, all possible minutiae pixels are detected by employing the corresponding direction extension vector. Then we get the minutia locations by computing the center of minutia pixels in a local region. Next, we get the minutiae directions by searching the largest extension distance. Finally, according to the relative location and relative direction between minutiae, false minutiae are eliminated. Experimental results show that the proposed approach can detect minutiae with less computational cost and lower false detection.

关 键 词：指纹 细节点检测 方向延伸向量 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]