Finger-vein recognition algorithm based on potential energy theory

Abstract

As a new biometric identification technology, finger-vein recognition technology is gathering more and more attention because it has some advantages such as: Non-contact, Live body identification, High security. In the finger-vein recognition system, there are many different characteristics matching algorithms after that the finger-vein vessel network has been segmented in the matching stage. In this paper, a new matching method based on potential energy is proposed for finger-vein recognition. This method first extracts the finger vein skeleton feature, and then uses a weighted potential energy vector of skeleton image to represent the skeleton feature, finally evaluate the similarity by calculating the distance of vectors to classify. Experimental results show that this method can make more effectively use of finger-vein vessel network position distribution information and achieve higher recognition accuracy.