Hyperspectral image classification via contextual deep learning

EURASIP Journal on Image and Video Processing

Table 2 Classification accuracy (%) for the Salinas image using 1 % training samples as shown in Fig. 6

Class	Train	Test	SAE-LR	LORSAL-MLL	SOMP	MPM-LBP	CDL-MLR
1	20	1989	100.0	100.0	99.90	100.0	100.0
2	37	3689	99.38	99.65	99.97	99.92	96.65
3	20	1956	92.61	95.56	92.77	96.20	99.95
4	14	1380	98.79	98.08	95.76	98.64	98.98
5	27	1651	97.86	96.82	99.49	98.64	98.78
6	40	3919	99.87	100.0	99.32	99.97	99.27
7	36	3543	99.38	99.94	100.0	99.94	100.0
8	113	11,158	83.32	80.10	86.61	88.64	96.62
9	62	6141	99.37	98.94	98.81	99.58	99.92
10	33	3245	96.07	94.54	95.21	92.27	98.57
11	11	1057	95.86	89.83	97.53	96.54	96.78
12	19	1908	99.48	95.11	99.95	98.27	98.37
13	9	907	94.84	95.30	97.73	96.47	98.58
14	11	1059	98.37	96.45	95.00	97.97	98.87
15	73	7195	78.34	78.28	71.31	88.83	95.64
16	18	1789	97.85	99.20	99.94	99.28	99.48
OA			95.12	91.56	92.00	95.12	98.26
AA			96.94	94.86	95.58	96.94	98.72
κ			0.9715	0.9058	0.9456	0.9456	0.9806