EURASIP Journal on Image and Video Processing

Table 3 Comparison of OA (%) and Kappa (%) on UC Merced Land-Use data set

From: Remote sensing scene classification based on rotation-invariant feature learning and joint decision making

Methods	Epoch	TP(%)
		20		50		80
		OA	Kappa	OA	Kappa	OA	Kappa
AlexNet	120	36.01	32.82	55.05	52.80	67.70	66.00
Siamese AlexNet	120	37.56	34.44	56.76	54.60	69.52	68.00
Siamese AlexNet+R.D	120	38.89	35.44	58.29	56.20	71.62	69.00
VGG16	120	76.53	75.56	85.05	84.30	90.24	89.75
Siamese VGG16	120	76.90	75.75	85.14	84.50	92.38	92.00
Siamese VGG16+R.D	120	78.13	76.94	88.10	87.50	93.33	93.00
ResNet50	120	74.11	72.81	89.43	89.00	91.90	91.50
Siamese ResNet50	120	76.52	75.06	90.95	90.50	94.29	94.00
Siamese ResNet50+R.D	120	78.87	77.81	91.71	91.20	94.76	94.50

Back to article page