摘要
针对目前指纹识别系统主要采用手指上细节点的分布来表征和匹配指纹,提出了一种采用指纹脊线特征的匹配算法,以提高细节点数量较少情况下的匹配精度.在特征提取阶段,通过脊线采样,只存储脊线采样点集以降低存储量;在匹配时,对欲匹配的两指纹利用细节特征配准脊线集,在重合区域内对两指纹脊线统一进行编码,通过编码的比较确定相似脊线;以相似脊线的相同位置编码为论域,以相同位置编码的相似程度为隶属度,建立衡量脊线相似程度的模糊集,采用加权平均法对多个相似脊线模糊集进行综合评判得到两指纹脊线总体相似度.最后将脊线匹配相似度与细节点匹配相似度进行加权融合得到两指纹最终的相似度.在FVC2004指纹库上的实验表明该算法能够有效提高指纹匹配的准确性.
Most fingerprint matching systems rely on the distribution of minutiae on the fingertip to represent and match fingerprints.This paper describes a matching scheme that used ridge flow information to represent and match fingerprints in order to improve the accuracy of fingerprints matching in case of a lack of sufficient minutiae.In the phase of features extraction,the ridges were sampled into sets of points so that the size of the template for storage could be shrunk significantly.In the matching phase,two fingerprints were aligned first,and the ridges in the overlap area were coded into sets of ridge codes.By comparing the ridge codes,similar ridge pairs were found.Taking the same position code of similar ridges as the domain,and the degree of similarity of the same position code as the degree of membership,the fuzzy sets of ridge similarity could be achieved.The ridge similarity of two fingerprints could be achieved through evaluating the fuzzy set using the weighted average method.Finally,the weighting fusion was made to calculate the final similarity of two fingerprints based on ridge matching similarity and minutiae matching similarity.Experiments show that this fingerprint matching technique can noticeably improve the accuracy.
出处
《智能系统学报》
北大核心
2012年第3期235-240,共6页
CAAI Transactions on Intelligent Systems
基金
辽宁省博士科研启动基金资助项目(20071066)
辽宁省教育厅科研计划资助项目(L2010072)
关键词
指纹识别
指纹匹配
模糊匹配
指纹脊线特征
指纹细节点特征
fingerprint identification
fingerprint matching
fuzzy matching
fingerprint ridge features
fingerprint minutiae features
