This paper presents a kernel weighted scatter difference discriminant analysis (KWSDA) method for face recognition. This nonlinear dimensionality reduction algorithm has several interesting characteristics. First, using a new optimization criterion it avoids small sample size problem intuitively. Second, by incorporating a weighting function into discriminant criterion, it overcomes overemphasis on well-separated classes and hence can work under more realistic situations. Lastly, applying kernel theory, it handles nonlinearity efficiently. Experiments on the ORL face database show that the proposed method is effective and feasible.
展开▼
