基于Matlab实现的指纹图像细节特征提取  

Minutiae Extraction of Fingerprint Image Based on Matlab

作　　者：郭晶莹[1] 吴晴[1] 商庆瑞[1] 

机构地区：[1]北京工业大学,北京100022

出　　处：《计算机仿真》2007年第1期182-185,共4页Computer Simulation

摘　　要：指纹图像的特征提取是指纹识别的关键,而指纹匹配通常基于细节点匹配。介绍了一套基于Matlab实现的指纹细节特征提取方法,并给出了去伪算法。指纹特征提取是从细化后的指纹图中得到细节特征点(即端点和分叉点),此特征点含有大量的伪特征,既耗时又影响匹配精度。采用了边缘去伪和距离去伪,使得特征点去伪前后减小了近1/3,然后提取可靠特征点信息,以便实现指纹匹配。实验证明,用Matlab实现的这种方法,既简单快速,而且具有较高的准确率。Minutiae extraction is essential in fingerprint recognition, and fingerprint match is used to search minutiae. This paper introduces a set of algorithms for extracting minutiae from fingerprint image based on Matlab. Minutiaes（ end minutia and bifurcation minutia） are extracted from thinned fingerprint image and they have a lot of false minutiaes. They are not only time - consuming, but also affect matching precision. Most false minutiae are eliminated by a method using edge - deleting and distance - threshold in this paper. Reliable minutiaes are extracted in order to achieve fingerprint match. The experimental results show that this algorithm of matlah is simple, quick and accurate.

关 键 词：指纹识别 细节点 特征提取 伪特征点 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]