A Binary Code Learning Method Based on Class-neighborhood for Finger Vein Recognition

Abstract

Abstract-The finger vein recognition, as a new biometric identification technology, judges the identity of a person by using their finger vein angiography images obtained by near-infrared light, and has recently drawn considerable attention from scientists worldwide. In this work, a novel binary code learning method based on class-neighborhood (BCLCN) is proposed for finger vein recognition. Different from traditional binary code methods based on pixels, BCLCN fully considers the relationships between categories of train set and encodes based those relationships. In this way, the discrimination of the code is enhanced and the binary codes are shortened. Experimental results on self-built finger vein image database demonstrate the effectiveness of BCLCN (EER = 0.512%).