Multispectral Palmprint Recognition by Feature Level Fusion

Abstract

Palmprint is a reliable and accurate trait in the biometric family. To improve the accuracy of the existing palmprint systems, multispectral imaging is used for palmprint recognition. Although image level fusion and matching score level fusion were widely used for multispectral palmprint recognition, feature level fusion was paid less attention. This paper proposes a novel method for multispectral palmprint recognition by the feature level fusion technique. Experimental results on a large public multispectral database show that the proposed method could get better results than image level fusion, and it could get comparable results with matching score level fusion with less cost of storage.