A New Kernel function to extract Non Linear Interval type Features Using Symbolic Kernel Fisher Discriminant Method with Application to Face Recognition

摘要

In this paper we propose to use a new RBF kernel function to extract non linear interval type features using symbolic kernel Fisher discriminant analysis (symbolic KFD) for face recognition. The kernel based methods are a powerful paradigm; they are not favorable to deal with the challenge of large datasets of faces. We propose to scale up training task based on the interval data concept. Our investigation aims at extending kernel Fisher discriminant analysis (KFD) to interval data using new RBF kernel function. We adapt the symbolic KFD to extract interval type non linear discriminating features, which are robust due to varying facial expression, view point and illumination. In the classification phase, we employed Euclidean distance with minimum distance classifier. The new algorithm has been successfully tested using three databases, namely, ORL database, Yale Face database and Yale Face database B. The experimental results show that symbolic KFD with new RBF kernel function outperforms other discriminant analysis based algorithms.