2D Pairwise FLD: A robust methodology for face recognition

In this paper a two dimensional pairwise Fisher's linear discriminant (FLD), 2DPairwise FED is proposed which is employed for face representation and recognition. The proposed methodology is robust as it is based on Pairwise-FLD which is theoretically more efficient than conventional FLD. In addition to this the proposed methodology has a higher recognition rate when benchmarked with contemporary models like 2D-PCA, Alt. 2D-PCA, 2D-LDA, Alt. 2D-FLD, 2D
²
-FLD and 2D-PCA during experimentation.