基于特征点融合小波能量特征的手背静脉识别

Dorsal Hand Vein Recognition Based on Wavelet Energy Feature Fused with Feature Points

手背静脉识别是生物识别领域的新兴课题,针对单一手背静脉识别方法在大量样本情况下正确识别率识较低的问题,提出了一种空间域特征融合小波域特征的识别方法,对预处理后的样本提取了空间域中的特征点以及在小波域中构造了小波能特征,并分别用改进的豪斯多夫距离以及加权城区距离进行度量,最后将两种方法进行加权融合,采用最近邻分类器进行识别;在具有100个样本的数据库上对该方法进行了测试,在最近邻分类阈值为9.46时识别率达到97.2%,表明了该方法的优越性。

Dorsal hand vein recognition is a new subject in the field of Biological recognition. For the problem of low recognition rate under one single dorsal hand vein recognition method in a large number of samples, a recognition method based on spatial domain feature com- bined with wavelet domain feature is proposed. We Extract the feature points in spatial domain and construct the wavelet energy feature in wavelet domain from the preprocessed samples. It used the improved Hausdorff distance and the weighted urban distance to measure, finally, the two methods are weighted fused and the nearest neighbor classifier is used for recognition. It is tested in the database with 100 samples, and the recognition rate is 97.2%0 in the condition of the nearest classification threshold is 9.46, so it shows the advantage of this method.