Face recognition using a hybrid algorithm based on improved PCA

Abstract

In order to reduce the heavy computation and eliminate the effects of uneven illumination in face recognition based on PCA (Principal Component Analysis) method, this paper proposed an improved PCA algorithm - A new hybrid face recognition algorithm which considered the image intensification based on local means and standard deviation, which was further integrated with genetic algorithm and form a new face recognition algorithm. Experimental results showthat,when tested byORLface database, the algorithm proposed in this paper is 3.27% higher than the PCA+2DPCA face recognition algorithm in average recognition rate. © 2014 Taylor & Francis Group, London.