Experimental Analysis: Hybrid Scheme for Face Recognition Using KPCA & SVD

For the various classification tasks of several visual phenomena, non-linear subspaces that derived from the kernel methods are preferable than the linear subspaces. In these methods some methods such as kernel principal component analysis, kernel singular value decomposition and kernel discriminant analysis are based on kernel approach. According to the studies and researches, incremental computation algorithms do not available also the practical implementation and execution of these methods on large database or online video processing is not at great extent. Here, we are experimentally discussing the hybrid scheme regarding integration of kernel principal component analysis and singular value decomposition algorithm. We have defined the steps of required algorithms involved in it and also the results from the experiments explore the efficacy of the suggested method.