Feature Fusion for Multiple Biometric Authentication

Abstract

Biometrics is the technology used for measuring and analyzing a person's unique characteristics. To overcome the problems in unimodal biometric authentication we present a multimodal approach based on feature level biometric fusion. We combine two kinds of biometrics: one is the palmprint feature and another one is fingerprint. The palm area contains a large number of features such as principal lines, wrinkles, delta points, minutia, and texture features. Fingerprints and palmprints have a unique structure formed by ridges, valleys and sweat pores present on the skin. Fingerprints have special features caused by ridge endings, bifurcations and loops. In this paper we propose a feature level fusion technique that extracts the discriminant features using gabor based image preprocessing and principal component analysis (PCA) from both palmprint and fingerprint. And then a statistical based normalized weighting strategy is used to conduct feature level fusion. Using PolyU palmprint and fingerprint database as the test data, the experimental results show that the presented approach significantly improves the recognition effect of single sample biometrics problem.