摘要
指纹图像增强和特征匹配是自动指纹识别系统的两个关键问题。但由于指纹在获取过程中各种干扰因素的存在,图像增强并不能完全消除噪音,会产生大量的伪特征点,从而降低匹配的速度及准确性.本文在指纹预处理的前提下,提出了一种有效的特征点的提取算法以消除伪特征点,从而保证特征点的稳定性、可靠性.
Finger image enhancement and minutiae matching are two key problems in Auto matched finger print identification system . But because of various disturbance in the process of acquiring fingerprint image,finger image enhancement can't eliminate noise completely,invalid minutiae are given rise to. In this paper a valid method of feature extraction is proposed from skeletonized binary image , so that the number of invalid minutiae is minimized and reliable minutiae are obtained.
出处
《南昌大学学报（工科版）》
CAS
2002年第4期38-41,共4页
Journal of Nanchang University(Engineering & Technology)
