Nested Transformers for Hyperspectral Image Classification

Table 3

Classification accuracy of different sampling rates for each dataset (%).

Salinas	Sampling rate
	1%	3%	5%
CNN	94.34	96.57	98.93
ViT	92.37	95.68	98.66
Swin	95.06	97.73	98.76
NesT	98.55	99.42	99.61
Indian Pines	Sampling rate
	1%	3%	5%
CNN	62.14	75.92	86.75
ViT	71.90	82.51	90.20
Swin	64.88	72.87	75.71
NesT	69.74	84.72	90.37
Tea Farm	Sampling rate
	1%	3%	5%
CNN	90.06	96.02	98.39
ViT	91.49	95.36	97.29
Swin	94.06	95.49	96.18
NesT	95.81	98.52	99.25
Xiongan	Sampling rate
	1%	3%	5%
CNN	98.45	98.93	99.70
ViT	98.32	99.42	99.73
Swin	99.26	99.79	99.90
NesT	98.31	99.52	99.80