基于连续分布方向图的指纹奇异点检测

Fingerprint Singularity Detection Based on Continuously Distributed Directional Image

在指纹连续分布方向图(场)的基础上,对经典的PoincaréIndex计算公式进行了改进,提出了一种新的基于连续分布方向图的指纹奇异点检测算法。由于指纹连续分布方向图过渡平滑、自然,既具有很好的连续性、渐变性和抗噪性,又具有较高的精确度;而改进后的PoincaréIndex不仅能精确表示向量场的旋转角度,而且还能精确表示向量场的旋转方向。所以,该算法能够在像素级水平精确定位指纹奇异点(core point和delta point),精确度达到一个像素。在FVC2000、FVC2002和FVC2004的训练指纹库(Set B)以及笔者采集的AFIS2004指纹库(含4000幅指纹)上的实验结果验证了该算法的有效性。

A new concept on the continuously distributed directional image/field（CDDF） and the method to compute it in the fingerprint images are proposed,which transits smoothly and exhibits not only good continuity,well gradualness and excellent robustness to noises,but very high precision,as well.Then,the classical formula to compute the Poincaré Index is improved so that it can present not only the rotation degrees,but also the rotation direction of the vector in the vector field,exactly.Finally,a novel method for fingerprint singularity detection based on both the continuously distributed directional image and the modified version of Poincaré Index in the fingerprint images is developed,which can not only locate the fingerprint singularities（core points,and delta points） at pixel level with an accuracy of only one pixel,but also extract only one candidate singular point（SP） within one true singular region（SR）.The experimental results obtained on the Set B fingerprint databases of the FVC2000,FVC2002 and FVC2004,and our live-scanned fingerprint database validate our algorithm and prove a substantial improvement in the locating accuracy and reliability of the fingerprint singularity detection.