基于方向场的半区域指纹分类方法

Classification approach of half-region fingerprint based on orientation field

提出了一种基于指纹方向场半区域指纹特征构造方法,并将此特征用于指纹分类.该方法充分利用了不同类型指纹方向场的结构特点,将指纹图像方向场中反映指纹全局特征的区域(感兴趣区域)分成许多小扇区,将扇区内所有像素方向的平均值作为该扇区的特征值,并将其按一定顺序排列起来构成一枚指纹的特征矢量.所提出的指纹特征矢量构造方法具有计算简单、特征矢量短等特点.实验结果表明,此方法构造的特征矢量在3种类型的指纹分类上具有令人满意的效果.

A new approach of fingerprint classification based on the interested region of the orientation field is described. In this approach, the interested region in fingerprint orientation field that can show the global information of fingerprint is divided into many small sectors, and then the average orientation values in the sectors are calculated. The feature vector of a fingerprint is formed by these average values in order. The proposed method has the characteristics of simple computation and short feature vector. The simulation results show that proposed algorithm is effective in the classification of three different types of fingerprints.