Local and Global Geometric Structure Preserving and Application to Hyperspectral Image Classification

Table 2

Classified performance for Indian Pine scene (%).

Item	Measurement
	Overall accuracy	Kappa coefficient	Average accuracy
1NN
PCA	67.37	62.72	71.02
MFA	28.57	19.25	26.51
LapLDA	48.29	40.96	50.68
RP	62.88	57.65	65.16
LPP	59.49	53.84	64.02
LGSPP	61.88	56.49	65.49
JGLDA	55.20	48.73	54.57
LGGSP	75.80	72.41	81.77
5NN
PCA	66.52	61.66	67.63
MFA	33.37	23.72	29.76
LapLDA	53.37	46.17	50.91
RP	64.34	58.99	65.21
LPP	60.98	55.13	62.35
LGSPP	63.63	58.19	64.36
JGLDA	58.51	51.84	51.47
LGGSP	76.62	73.33	80.99
9NN
PCA	66.32	61.26	67.27
MFA	36.14	25.98	29.39
LapLDA	55.66	48.32	49.08
RP	63.15	57.50	61.61
LPP	60.45	54.40	59.32
JGSPP	62.98	57.21	63.44
JGLDA	59.63	52.70	49.52
LGGSP	75.75	72.28	79.31
Linear SVM
PCA	57.13	48.55	52.00
MFA	36.93	20.20	16.92
LapLDA	50.89	40.82	28.52
RP	52.32	41.90	31.22
LPP	54.47	45.59	46.36
JGSPP	54.87	45.35	41.65
JGLDA	47.88	35.26	33.52
LGGSP	63.08	55.95	48.33
Polynomial SVM
PCA	62.40	55.74	60.29
MFA	41.78	30.94	29.77
LapLDA	51.97	43.37	37.87
RP	55.14	45.60	43.32
LPP	58.83	51.67	55.61
LGSPP	60.05	52.66	57.36
JGLDA	58.25	50.49	51.65
LGGSP	72.64	68.14	72.57
RBF-SVM
PCA	73.69	69.92	76.21
MFA	24.26	0.12	6.35
LapLDA	56.31	48.94	50.06
RP	71.90	67.76	72.54
LPP	68.45	63.90	71.64
LGSPP	67.33	62.33	66.76
JGLDA	54.79	49.46	58.71
LGGSP	79.04	75.86	76.95