针对掌纹纹理特点,提出一种基于灰度曲面匹配的掌纹识别方法。将掌纹灰度图像看作是三维空间下的灰度曲面,并将该曲面作为特征进行匹配,由于将整个图像作为特征,有难以被伪造的特点,抛弃了特征提取和编码等传统操作,提高了掌纹识别系统的安全性和识别速度。对100个手掌类的1 000幅掌纹图像数据库,进行了匹配实验。用Haar小波降低分辨率到3级的情况下,其正确识别率仍然能够达到98.45%。识别时间为23.5 ms。理论分析与实验结果表明,该方法识别准确率高,识别速度快。
Corresponding to the feature of palrnprint,a recognition method is proposed based on the gray surface match, which considers the palmprint gray image as a gray surface in 3D space and uses it as feature to match. This method can't be fabricated,and deserts the traditional operation like extracting features and coding, so it enhances the security and speed of the recognition system. Matching experiments are carried on a database of 1000 images of 100 palms and the correct recognition rate can reach 98.45 when the resolution are lowered into 3rd level using Harr wavelet. Theory analysis and experiment resuits show that this method achieves high recognition rate and speed.