Improved PCA facial recognition with bootstrap and data standardization in small sample case
Shunfang Wang, Fu-Lai Ku, Huai-Xiong Zhang · 2011
Based on the traditional principal component analysis (PCA) in facial recognition, first this paper proposes an bootstrap resampling method in PCA to overcome the small sample problem in the process of data mean centering. Second this paper uses the data standardization to improve the mean centering process. Experiments suggest that the proposed method is much better than the traditional PCA in facial recognition.