基于区域分割的指纹奇异性检测及中心点计算  

An algorithm for singularity detection and center point calculation based on fingerprint segmentation

作　　者：乌旭[1] 胡家升[1] 梁殿亮[1] 

机构地区：[1]大连理工大学电子与信息工程学院

出　　处：《光学精密工程》2006年第2期229-235,共7页Optics and Precision Engineering

摘　　要：提出了一种基于区域分割的指纹奇异性检测以及中心点计算算法。该算法在对指纹奇异性分析的基础上,采用Gabor滤波、方差分析、图形二值化等分析手段,将指纹图样按不同的指纹脊线方向进行区域分割,根据合成图样每个元素邻域的灰阶变化次数以及相应的灰阶信息来检测指纹的奇异区。最后通过计算各奇异区质心的方法,实现指纹奇异点的定位,并进一步讨论了通过指纹奇异点计算指纹中心点的方法。实验结果表明,本算法可以有效实现不同类型指纹奇异性的检测及中心点的准确定位,算法简单,具有较好的鲁棒性。A novel and simple algorithm for singularity detection and center point calculation in a fingerprint image was presented. Based on analysising of the property of singular points, the Gabor filter,variance analysis were introduced to segment the fingerprint on ridge orientation effectively. By the binary image addition of the fingerprint segmentation, the singular regions were detected according to the change times of gray levels and the intensity in each pixel＇s neighborhood. The calculation of singular points and center point of the same fingerprint was discussed. The experimental results show that the proposed algorithm can detect more sorts of singular points than other conventional methods and can realize the location of the singular point and center point correctly.

关 键 词：方向滤波 GABOR滤波 指纹增强 奇异点 中心点 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]