An extended PCA and LDA for color face recognition

This paper proposes a new method for color face recognition that extracts the first, second and third channels of each color image and uses PCA and LDA to get a score of each channel, then uses a combine scheme to attain a final score and uses this final score to classify test samples.

Abstract

In this paper, we propose a new method for color face recognition. We extract the first, second and third channels of each color image and use PCA and LDA to get a score of each channel, then we use a combine scheme to attain a final score and use this final score to classify test samples. In order to check the performance of our method, we conduct experiments on Georgia Tech (GT) color face database, at the same time, we compare our method with PCA and LDA, and experiment results show that our methods take better performance.