Gabor-based weighted region covariance matrix for face recognition

Abstract

A novel image region descriptor, named the Gabor-based weighted region covariance matrix (GWRCM), is proposed by incorporating the weighting matrices into a Gabor-based region covariance matrix (GRCM). Based on the weighting matrix, the proposed approach can assign different weights for pixels in different face parts, so it preserves more discriminating information than the GRCM for face recognition. Experimental results show that the proposed GWRCM approach outperforms other algorithms in terms of recognition accuracy.