Table 10 Overall accuracy (%) of different architectures on NWPU-RESISC45 dataset.
From: A multi-scale dense residual correlation network for remote sensing scene classification
Scheme | Architecture | OA(10/90)(%) | OA(20/80)(%) |
|---|---|---|---|
1 | Without MSFE | 90.12\(\,\pm \,\)0.25 | 91.86\(\,\pm \,\)0.26 |
2 | Without DRCFF | 91.09\(\,\pm \,\)0.34 | 92.61\(\,\pm \,\)0.41 |
3 | Without CAM | 91.39\(\,\pm \,\)0.47 | 92.98\(\,\pm \,\)0.26 |
4 | Ours | 91.59\(\,\pm \,\)0.29 | 93.82\(\,\pm \,\)0.17 |
