Applying Weighted K-nearest centroid neighbor as classifier to improve the finger vein recognition performance

Recently, finger vein recognition technology, which works based on physiological characteristics of finger vein patterns, has been widely developed as the most promising biometric technology due to the excellent advantages in application such as uniqueness, universality, highest performance and measurability. In this article, we proposed a new algorithm for finger vein recognition combining of Kernel principal component Analysis (KPCA) and a new effective classifier called Weighted K-nearest centroid neighbor (WKNCN) in order to improve the finger vein recognition performance. Experimental results demonstrate that the proposed algorithm obtains much improvement in pattern recognition.