基于多峰检测和能量权重的指纹纹线距离估计方法  

Ridge distance estimation methods based on multi-peak detection and energy weight

作　　者：张晓梅[1] 贾同辉[1] 尹义龙[1] 詹小四[1] 

机构地区：[1]山东大学计算机科学与技术学院,山东济南250100

出　　处：《山东大学学报（理学版）》2008年第3期34-39,共6页Journal of Shandong University(Natural Science)

基　　金：山东省优秀中青年科学家科研奖励基金资助项目(2006BS01008);山东省高新技术自主创新创新工程基金资助项目(2007ZCB01030);山东省科技攻关计划基金资助项目(2005GG3201089);山东省博士后基金资助项目(2000602014)

摘　　要：指纹纹线距离是指纹的本质属性,也是目前主流指纹增强技术中的常用参数,而指纹增强在指纹识别过程中起着去除噪声影响并同时突出关键特征信息的重要作用,因此,纹线距离估计是自动指纹识别系统的一个关键环节。基于精确、可靠地实现指纹纹线距离估计仍然是一个比较困难的任务,提出了基于多峰值检测和基于能量权重2种方法实现纹线距离的估计,并通过在指纹图像增强和指纹细节点提取过程中的应用对纹线距离效果估计进行验证。实验结果表明,该方法可以获得较好的纹线距离估计效果。The ridge distance of fingerprints is an essential attribute of fingerprints, and an important parameter used in popular fingerprint enhancement algorithms. Fingerprint enhancement has important functions in fingerprint identification, such as reducing noise and emphasizing key features. As a result, ridge distance estimation is a key step in AFIX （automated fingerprint identification system）. However, it is still a difficult task to accurately and reliably estimate the ridge distance. Two methods for estimating the ridge distance, namely, the method based on multi-peak detection and the method based on energy weight, were put forth. In addition, the performance of these two methods is verified by applying ridge distance estimation results to fingerprint enhancement and minutiae extraction. Experimental results indicate that these two methods have a good impact on ridge distance estimation.

关 键 词：指纹 指纹识别 指纹增强 纹线距离估计 频谱分析 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]