Abstract
Palm vein recognition is a promising biometric technology, it uses vascular pattern as personal identification data. In this paper, we propose a new algorithm of palm vein recognition and use local binary pattern (LBP) matching strategy to conduct experiments. First, we discuss two methods to obtain ROI of palm vein. Then the maximal principal curvature (MPC) algorithm and k-means method are utilized to extract the features of palm vein. Finally, template matching and LBP are used to recognize the feature. A series of experiments on CASIA multispectral palm image database were conducted. The lowest equal error rate (EER) is 0.01965. And relative side length could extract more useful information than fixed side length.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
