基于复数滤波的指纹奇异点提取方法

在身份识别中,指纹具有唯一性,人的指纹包含大量信息,识别指纹图像很重要.针对指纹图像奇异点中准确判断和精确定位的难题,提出了一种改进的指纹奇异点提取方法.首先,对指纹平方复数点方向场进行多尺度滤波,在平滑噪声的同时,保留奇异区方向的细节信息.对方向场进行复数滤波,并运用启发式规则增强复数滤波的响应幅度,通过分析响应的幅度检测奇异点的位置,由响应的相位确定方向.在FVC 2000上进行实验.结果表明,上述方法可精确提取指纹图像中各奇异点的位置及其方向信息,证明优于与其它方法,结果具有较好的鲁棒性.

An Extraction Method of Fingerprint Singular Points Based on Complex Filtering

Fingerprint singular points have important applications in fingerprint classification and feature matc-hing. An improvod method of fingerprint singular points extraction is proposed to improve the accuracy of the position and reliability of the singular points. First, the method executes multiple scale filtering on fingerprint square complex orientation field in pixel-wise resolution, at the same time, the noise is smoothed and singular area details are re-tained, then takes complex filtering on orientation field coming with several heuristic rules to enhance the response amplitudes. Finally, it detects the position of singular points by analysing the magnitude of response and determines their direction by the phase response of singular points loeation. Experimental results of FVC 2000 show that the method can accurately locate singular points and determine their orientations in fingerprint images, comparison with other methods also shows better effectiveness and robustness.