An Improved EMD-Based Dissimilarity Metric for Unsupervised Linear Subspace Learning

Table 17

The recognition rates on noise Extended Yale B face database over “F1” and “F2” features; recognition rates (dimensionality).

Face database	Algorithms	G6/P10	G7/P9	G8/P8	G9/P7	G10/P6
Noise	Baseline (F1)	0.2623 (32256)	0.2750 (32256)	0.2833 (32256)	0.3057 (32256)	0.3119 (32256)
	LPP (F1)	0.4345 (105)	0.4939 (115)	0.5123 (120)	0.5443 (135)	0.5839 (140)
	Algorithm 1 (F1)	0.5152 (105)	0.5663 (115)	0.5758 (130)	0.6055 (125)	0.6350 (95)
Noise	Baseline (F2)	0.2435 (360)	0.2496 (360)	0.2654 (360)	0.2821 (360)	0.2881 (360)
	LPP (F2)	0.1802 (125)	0.1963 (140)	0.2031 (140)	0.2131 (150)	0.2283 (170)
	Algorithm 1 (F2)	0.2530 (95)	0.2604 (120)	0.2702 (145)	0.2795 (160)	0.2917 (165)