Multiple-Biometric Fusion Methods using Support Vector Machine and Kernel Fisher Discriminant
Nayoung Woo, Hakil Kim · 2006
This paper proposes multiple-biometric fusion methods, using Support Vector Machine (SVM) and Kernel Fisher Discriminant (KFD) in order to improve recognition accuracy. The proposed SVM and KFD are non-linear classification methods using Radial Basis Function (RBF) as a kernel function. The RBF kernel is effective with respect to the input data distribution. Experiments have been conducted on NIST BSSR1 (Biometric Scores Set - Release1) data set using SVM, KFD and weighted sum methods, and their performance of multiple biometric fusion has been presented by the HTER (Half Total Error Rate) and the ROC (Receiver Operating Characteristic) curves. The experimental results demonstrate that the non-linear fusion methods provide higher verification performance than linear fusion methods.