An Improved Kernel Feature Extraction Method and its Application to Face Recognition

The kernel trick is used firstly to project the original samples into an implicit space called feature space by nonlinear kernel mapping, then two equivalent models based on Fisher discriminant minimal criterion are established by the theory of reproducing kernel in the feature space.Finally, Fisher discriminant minimal criterion is carried out in the null space and non-null space going between kernel-class scatter of feature space, to obtain the optimal kernel discriminant vectors.Experimental results on face database show the effectiveness of the algorithm proposed.