Improving the performance of color spaces for face recognition by color space normalization and rotation

Recent research efforts show that color provides useful information for face recognition. Color face representation and recognition is necessarily related to a color space. This chapter presents a color space normalization and rotation (CSNR) technique for enhancing the discriminating power of color spaces for face recognition. Different color spaces usually display different discriminating power, and our experiments on a large scale Face Recognition Grand Challenge (FRGC) problem reveal that the RGB and XYZ color spaces are weaker than the I₁I₂I₃, YUV, YIQ, and LSLM color spaces for face recognition. We therefore apply our CSNR technique to the weak color spaces, such as the RGB color space, the XYZ color space and the three hybrid color spaces XGB, YRB and ZRG. Experimental results using the most challenging FRGC version 2 Experiment 4 with 12,776 training images, 16,028 controlled target images, and 8,014 uncontrolled query images, show that the proposed CSNR technique can significantly improve the discriminating power of color spaces. In particular, the CSNR - generated color spaces are demonstrated more powerful than the I₁I₂I₃, YUV, YIQ and LSLM color spaces. We further explain why the CSNR technique can improve the recognition performance of color spaces from the color component correlation point of view.