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

Table 23

The recognition rates on noise Extended Yale B face database over “F3” features; recognition rates (dimensionality).

Face database	Algorithms	G6/P10	G7/P9	G8/P8	G9/P7	G10/P6
Noise	Baseline (F3)	0.4442 (5760)	0.4598 (5760)	0.4746 (5760)	0.5043 (5760)	0.5131 (5760)
	LPP (F3)	0.2923 (105)	0.2893 (185)	0.3075 (180)	0.3217 (210)	0.3428 (210)
	Algorithm 2 (F3)_	0.4397 (80)	0.4613 (115)	0.4890 (130)	0.5205 (210)	0.5350 (245)
	Algorithm 2 (F3)	0.4437 (150)	0.4419 (165)	0.4648 (215)	0.4864 (215)	0.4950 (235)