Optimal Decision Fusion and Its Application on 3D Face Recognition

Abstract

Fusion is a popular practice to combine multiple classifiers or multiple modalities in biometrics. In this paper, optimal decision fusion (ODF) by AND rule and OR rule is presented. We show that the decision fusion can be done in an optimal way such that it always gives an improvement in terms of error rates over the classifiers that are fused. Both the optimal decision fusion theory and the experimental results on the FRGC 2D and 3D face data are given. Experiments show that the optimal decision fusion effectively combines the 2D texture and 3D shape information, and boosts the performance of the system.