提出一种具有信息丰富、稳定性强、低分辨率可采集等特征的掌纹纹理识别方法。该方法首先从6个方向计算掌纹图像每点的相位一致性,把该点相位一致性最大值对应的方向作为对数Gabor滤波器的方向,在此方向上用对数Gabor滤波器提取相位信息,并利用实部和虚部进行量化和编码掌纹相位象限码,码长为256字节。对100个手掌类,共1000幅掌纹图像的数据库进行实验,在给定阈值的情况下,识别率能达到99.95%左右。把实验结果与典型的基于掌纹纹理特征的识别算法进行比较,表明该方法正确识别率高。

A method of recognition of the palmprint texture,with rich information,strong stability,and low resolution collection is proposed. This method first calculates the phase congruence of every point in the palmprint image from 6 directions and sets the direction corresponding to the maximum of phase congruence of every point as the direction of log-Gabor filter,and then at that direction extracts phase information of every point of the image with log-Gabor filter,and further quantifies and codes real and imaginary parts of the phase information. Thus the palmprint phase quadrant code with 256 bytes in length is obtained. Experiments are carried on the database of 1000 images of 100 palms. The result shows that the recognition rate could reach 99.95% at given threshold. Compared with the traditional recognition algorithm based on the palmprint texture feature,this method has higher recognition rate.