Two Dimensional Principal Component Analysis based Independent Component Analysis for face recognition
Xingfu Zhang, Xiangmin Ren · 2011
We usually reduce the dimensionalities of the data before running many algorithms of processing images and audio. Then we can remove the redundant data and reserve the useful features for future analysis. Independent Component Analysis is a famous dimensionality reduction algorithm. We usually run Principal Component Analysis algorithm firstly as a preprocessing procedure for decreasing the computation complexity before running Independent Component Analysis algorithm. We proposed Two Dimensional Principal Component Analysis based Independent Component Analysis algorithm, which processed the two dimensional images directly in preprocessing procedure. The contrast experiments on Yale databases prove that our algorithm is more effective than classical PCA, 2dPCA and ICA algorithms.