Table 5 Computational efficiency comparison of HSICNet with baseline models.

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

Model

Params (Millions)

Training time/Epoch (s)

Inference time/Image (ms)

FLOPs (GigaOps)

3D CNN

3.8

12.5

18.4

9.6

ResNet18

11.2

14.8

22.7

12.3

3D CNN + SE Attention

4.6

13.9

19.8

10.5

HSICNet

2.1

7.6

9.3

4.2

Back to article page