摘要
误识率和拒识率是指纹自动识别技术中非常重要的性能指标。使用一种新的基于局部结构和全局结构的指纹细节点匹配算法可以降低指纹自动识别的误识率和拒识率,该算法先对细节点间的相对距离作归一化处理,然后利用细节点的局部2-邻域结构不变性对指纹进行初步匹配,最后再根据细节点的全局结构进一步匹配指纹,提高指纹匹配的准确性。实验结果表明该算法能够有效降低误识率和拒识率。
FAR and FRR are the most important performance indexes in automatic fingerprint technology.This paper applies local and global structures of fingerprint minutiae to mach fingerprint images to reduce FAR and FRR.The algorithm firstly normalizes the ridge frequency to decrease the noise,then utilizes the 2-neighborhood local structure of minutiae to match images and gets the best-matched minutiae pair,last matches images according to the global structure of minutiae in advance.Experimental results show the excellent performance of the proposed algorithm and effective reduction of FAR and FRR.
出处
《浙江理工大学学报(自然科学版)》
2011年第5期758-762,共5页
Journal of Zhejiang Sci-Tech University(Natural Sciences)
关键词
指纹
细节点
局部结构
全局结构
匹配
fingerprint
minutiae
local structure
global structure
match