Discriminative Label Relaxed Regression with Adaptive Graph Learning

Table 6

Average recognition rate of different methods on the UMIST database (%).

Methods	The number (or proportion) of the training samples per class
	1 (21.05%)	2 (26.31%)	3 (31.58%)	4 (36.84%)	5 (42.10%)
CRC	46.77	60.00	71.55	74.68	84.33
LRC	48.53	60.56	71.72	75.63	84.60
FLRC	49.93	60.88	72.87	76.24	83.97
FME	56.25	58.28	69.79	75.31	83.29
JGLDA	53.13	59.67	69.66	78.59	87.80
DLSR	55.45	60.00	74.25	80.23	85.98
DLRR_AG	56.78	61.69	73.79	81.61	88.28

Bold values indicate the highest recognition rate of all methods.