基于宏观曲率的指纹特征提取和分类

特征提取是在自动指纹识别系统中重要的一步。为了有效地利用大量的指纹弯曲信息对指纹进行描述，提出了一种新算法。该算法通过综合邻近几条指纹脊线的信息，求出了一种能反映指纹宏观弯曲规律，并对单条指纹曲线的不规则不敏感的特征。随后提出了一种基于这些特征的指纹分类方法，实验结果表明，该算法提取出的特征清楚的描述了 纹脊线的弯曲规律，同时对噪声不敏感，可作为指纹识别的辅助特征。新的分类方法可以减少指纹匹配的搜索空间。

Fingerprint Feature Extraction and Classification Based on Macroscopic Curvature

In an automatic fingerprint identification system (AFIS), feature extraction is a critical step For effectively using the curve information to describe fingerprint, a novel algorithm is proposed; it embraces information of few fingerprint ridges nearby to extract a new characteristic which can describe the curvature feature of fingerprint Furthermore a new classification method based on macroscopic characteristics is proposed. Experimental results demonstrate that the algorithm is feasible, and the characteristics extracted by it can clearly show the inner macroscopic curve properties of the fingerprint image The result also shows that this kind of characteristic is robust to noise and pollution, e g those from additive Gaussian noise, and can be applied in fingerprint verification as a supplement to the customary minutiae The searching space in matching process is greatly reduced with the aid of the new classification method