基于方向图的指纹奇异点快速检测方法

摘要

It is significant to classifying and matching fingerprints that to detect singularities ( core points and delta points) accurately and quickly. In this paper, an algorithm for calculating fingerprint directional image is proposed first. Then, based on the directional image, we improve classical algorithm of fingerprint singularity detection which uses Poincare-based index computation formula. At last, the improved algorithm is employed to simulate on Matlab 1000 fingerprint samples collected by PU -JY203U police-used live fingerprint collection instrument. The result shows that compared with classical algorithm, the improved one is meliorated a lot in exactly and quickly locating the singularities.%准确、快速地检测指纹奇异点(core点和delta点),对指纹分类和指纹匹配等具有重要意义.首先,给出计算指纹图像方向图的算法,然后,在基于方向图的基础上,对传统的基于poincare索引计算公式的指纹奇异点检测算法进行了改进.使用改进算法对PU-JY203U警用活体指纹采集仪采集的指纹(1000枚)进行Matlab仿真.实验结果表明改进后的算法与传统算法相比,在定位奇异点的精确度和速度上都有很大的提高.