改进的主曲线算法在指纹骨架提取中的应用  

Improvement of principal curves algorithm and its application in finger-print skeleton extraction

作　　者：马驰[1,2] 张红云[3] 苗夺谦[3] 

机构地区：[1]北京科技大学经济管理学院,北京100083 [2]辽宁科技大学计算机科学与工程学院,辽宁鞍山114051 [3]同济大学计算机科学与技术系,上海200092

出　　处：《计算机工程与应用》2010年第16期170-173,共4页Computer Engineering and Applications

基　　金：国家自然科学基金(No.60775036;No.60703007)~~

摘　　要：在指纹识别系统中,指纹骨架的提取是一个重要而困难的环节,传统的细化算法得到的骨架很容易受到噪声的干扰。因此,在研究指纹数据特点和Kégl的主曲线算法的基础上,提出了一种改进的主曲线算法。实验结果表明:改进算法与传统算法相比有着更好的效率和效果,它所提取的指纹骨架包含更多的信息且具有更高的准确性、可靠性和抗噪声性。In the fingerprint recognition system,skeleton extraction for fingerprint images is an important and difficult task.Traditional thinning algorithms are susceptible to noise.In view of this,according to some characteristics of the fingerprint dataset and study on the original principal curves algorithm proposed by Kégl,an improved principal curves-based algorithm is proposed,which can be served as the skeleton of a fingerprint.Experimental results show that the improved principal curve algorithm is better in efficiency and quality than the original algorithm.The skeleton obtained by improved algorithm contains more information quantity and is proved to be more accurate,reliable and anti-noisy.

关 键 词：主曲线 指纹 细化 骨架提取 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]