基于细节点频谱的指纹图像匹配算法研究

Research of Fingerprint Image Matching Algorithm Based on Spectral Minutiae

为了提高指纹识别系统对存在位移、缩放和旋转等线性变化的指纹图像识别率,提出一种基于指纹细节点频谱的指纹匹配算法。首先介绍了从频率域提取指纹细节点的频谱特征,利用傅里叶-梅林变换的特性来消除平移、缩放和旋转等线性变化,从而获得稳定的无需再次校准的指纹特征;由于新的频谱特征是以矩阵形式存储的,因此可以直接用欧氏距离来评估两个指纹特征的相似程度,进而完成指纹图像的匹配;最后利用Matlab搭建了一个系统平台,在FVC2002DB2数据库中进行实验。结果表明:与其他方法相比,笔者提出的方法明显提高了自动指纹识别系统的识别率。

In order to improve the recognition rate of fingerprint images with displacement,rotation and scale,this paper proposes fingerprint matching algorithm based on the spectral of fingerprint minutiae.Firstly,this paper extracted the feature of fingerprint minutiae in frequency domain.Secondly,this paper used the characteristics of Fourier-Mellin transform(FMT)to eliminate the linear changes(such as translation,scaling and rotation)to obtain the alignment-free and stable fingerprint feature.Since the new fingerprint minutiae spectral feature was stored in the form of matrix,Euclidean distance could be directly used to evaluate the similarity degree of two fingerprint features so as to complete the fingerprint image matching.Finally,this paper utilized matlab to build a system platform and carried out experiment in the FVC2002DB2 database.The experimental result shows that:compared with other methods,the proposed method significantly improves the precision of the automatic fingerprint recognition system.