Fusing Two Kinds of Virtual Samples for Small Sample Face Recognition

Table 6

Classification error rates of different combinations of weights on the Yale database.

Different combinations of weights	Classification error rates
= 1, = 0, = 0, = 0 (original)	0.1083
= 0.7, = 0.3, = 0, = 0 (mirror face)	0.0333
= 0.6, = 0, = 0.2, = 0.2 (symmetrical face)	0.0500
= 0.5, = 0.1, = 0.2, = 0.2	0.0250
= 0.5, = 0.2, = 0.15, = 0.15	0.0333
= 0.6, = 0.1, = 0.15, = 0.15	0.0417
= 0.7, = 0.1, = 0.1, = 0.1	0.0500
= 0.8, = 0.1, = 0.05, = 0.05	0.0500
= 0.9, = 0.05, = 0.025, = 0.025	0.0833