Spectral-Spatial Hyperspectral Image Semisupervised Classification by Fusing Maximum Noise Fraction and Adaptive Random Multigraphs

Table 5

Classification results of different techniques for the Indian Pines dataset (10 classes).

No.	Pixon-based	PCA-SPCA-2D-SSA	R-VCANet	RN-FSC	iCapsNet	RPNet	MBFSDA	SS-MNF-ARMG
1	41.42	61.55	54.94	72.29	64.71	60.32	70.57	75.38
2	33.93	86.49	74.09	78.97	95.56	73.38	81.06	95.6
3	71.43	99.32	83.54	90.59	82.09	93.56	81.09	100.00
4	23.16	96.05	94.78	89.82	99.06	84.2	82.06	93.2
5	87.73	99.37	96.13	99.79	99.59	96.73	99.39	99.79
6	65.81	87.97	75.55	86.61	73.35	86.47	77.48	93.53
7	76.09	86.28	61.43	89.94	88.01	81.48	66.37	92.43
8	22.96	97.01	80.64	98.25	74.59	60.1	73.94	97.69
9	95.44	98.85	88.15	82.61	98.84	58.73	90.8	97.63
10	91.58	93.52	82.63	94.82	96.84	95.79	59.21	94.82
AA	60.96	90.64	79.19	88.37	87.27	79.08	78.2	93.01
OA	62.45	87.27	86.28	86.16	85.83	75.94	76.42	91.05
Kappa	56.7	85.3	72.05	83.97	83.63	72.32	72.95	89.61

Bold values represent the best results among these methods.