Table 4 Performance comparison of the proposed HSICNet with recent deep learning models.

From: HSICNet a novel deep learning architecture for hyperspectral image classification in remote sensing and environmental monitoring

Model

Dataset

OA (%)

AA (%)

Kappa

F1-score

3D CNN

Indian Pines

94.23

93.11

0.92

0.927

ResNet18

Indian Pines

95.60

94.75

0.94

0.943

3D CNN + SE Attention

Indian Pines

96.83

95.69

0.96

0.962

HSICNet

Indian Pines

98.63

97.42

0.98

0.976

3D CNN

Pavia University

96.74

95.88

0.95

0.956

ResNet18

Pavia University

97.21

96.11

0.96

0.961

3D CNN + SE Attention

Pavia University

97.95

97.07

0.97

0.972

HSICNet

Pavia University

99.14

98.30

0.99

0.986

3D CNN

Salinas

97.11

96.44

0.96

0.964

ResNet18

Salinas

97.45

96.73

0.96

0.967

3D CNN + SE Attention

Salinas

98.20

97.81

0.98

0.979

HSICNet

Salinas

99.35

98.91

0.99

0.991

Back to article page