A kernel fractional-step nonlinear discriminant analysis for pattern recognition

Feature extraction is one of the most significant and fundamental problems in pattern recognition (PR). This paper introduces a novel kernel fractional-step nonlinear discriminant analysis (KF-NDA) for feature extraction in PR. It not only overcomes the limitation of failing for a nonlinear problem in the direct fractional-step linear discriminant analysis (DF-LDA), but also improves the generalization ability of traditional kernel nonlinear discriminant analysis (K-NDA). It is then applied to an experiment on face recognition, and the results demonstrate that this method is more effective than the existing methods.