Table 8 OA results across various methods on the NWPU45 dataset.
From: Pure data correction enhancing remote sensing image classification with a lightweight ensemble model
Method | Technique feature | Params (M) | Published year | OA (%) | |
|---|---|---|---|---|---|
TR-10% | TR-20% | ||||
TPENAS-CNN6 | Single CNN | 1.8 | RS2023 | None | 90.38 |
SLGE-CNN7 | 5.1 | TGRS2022 | 96.44 ± 0.21 (TR-80%) | ||
AF-CNN8 | 3.8 | TGRS2022 | 95.32 (TR-60%) | ||
LCNN-HWCF11 | 0.6 | RS2022 | 93.10 ± 0.12 | 94.53 ± 0.25 | |
SCCNN42 | 0.5 | RS2022 | 92.02 ± 0.50 | 94.39 ± 0.16 | |
MF2CNet43 | 33.2 | TGRS2022 | 92.07 ± 0.22 | 93.85 ± 0.27 | |
ViT-Base44 | Single ViT | 86.6 | RS2021 | 93.83 ± 0.46 | None |
ET-GSNet12 | 98.3 | TGRS2022 | 92.72 ± 0.28 | 94.50 ± 0.18 | |
ViTAEv213 | Single Swin-T | 27.6 | TGRS2023 | 94.41 ± 0.11 | 95.60 ± 0.06 |
MBLANet46 | Attention module for CNN | > 25.6 | TIP2022 | 92.32 ± 0.15 | 94.66 ± 0.11 |
EAM-CNN14 | > 46.8 | GRSL2021 | 91.91 ± 0.22 | 94.29 ± 0.09 | |
LHNet47 | Feature fusion | > 46.8 | TGRS2022 | 89.89 ± 0.15 | 92.53 ± 0.13 |
SCViT15 | 40.1 | TGRS2022 | 92.72 ± 0.04 | 94.66 ± 0.10 | |
MLF2Net48 | 23.8 | GRSL2022 | 92.35 ± 0.17 | 94.84 ± 0.09 | |
SEMSDNet49 | 3.7 | JSTARS2021 | 91.68 ± 0.39 | 93.89 ± 0.63 | |
LmNet50 | > 25.0 | ACESS2021 | 93.00 ± 0.11 | 94.85 ± 0.14 | |
MLFC-Net64 | Multiple models | 65.2 | CG2022 | 92.52 ± 0.38 | 94.76 ± 0.08 |
D-CNN51 | None | RS2021 | 89.88 | 94.44 | |
GCSANet16 | 8.1 | JSTARS2022 | 93.39 ± 0.39 | 94.95 ± 0.36 | |
SF-MSFormer18 | 36.3 | TGRS2023 | 92.74 ± 0.23 | 94.83 ± 0.13 | |
AGOS52 | > 12.5 | TGRS2022 | 93.04 ± 0.35 | 94.91 ± 0.17 | |
MGSN65 | > 12.0 | JSTARS2022 | 91.92 ± 0.12 | 94.33 ± 0.08 | |
GRMA-Net19 | 54.1 | TGRS2022 | 93.67 ± 0.21 | 95.32 ± 0.28 | |
ACNet53 | > 276.6 | JSTARS2021 | 91.09 ± 0.13 | 92.42 ± 0.16 | |
T-CNN54 | 15.9 | TGRS2022 | 90.25 ± 0.14 | 93.05 ± 0.12 | |
GLDBS55 | > 23.4 | GRSL2022 | 92.24 ± 0.21 | 94.46 ± 0.15 | |
TRS56 | 46.3 | RS2021 | 93.06 ± 0.11 | 95.56 ± 0.20 | |
CTNet57 | > 107.8 | GRSL2022 | 93.90 ± 0.14 | 95.40 ± 0.15 | |
HHTL58 | > 173.2 | JSTARS2022 | 92.07 ± 0.44 | 94.21 ± 0.09 | |
P2FEViT20 | > 112.2 | RS2023 | 94.97 ± 0.13 | 95.74 ± 0.19 | |
L2RCF20 | Custom learning framework | 46.7 | TGRS2023 | 94.58 ± 0.16 | 95.60 ± 0.12 |
ViT-CL61 | 86.0 | JSTARS2023 | 92.85 | 94.69 | |
GSCCTL62 | None | IJRS2022 | 91.96 | None | |
MGDNet66 | None | TGRS2023 | 84.81 ± 0.36 | 91.41 ± 0.69 | |
LGRIN67 | 4.6 | TGRS2022 | 91.91 ± 0.15 | 94.43 ± 0.16 | |
TSTNet21 | Multiple Swin-Ts | 173.0 | RS2002 | 94.08 ± 0.24 | 95.70 ± 0.10 |
IBSwin-CR22 | 164.0 | JSTARS2023 | 93.98 ± 0.24 | 95.65 ± 0.11 | |
MFST63 | None | GRSL2022 | 92.64 ± 0.08 | 94.90 ± 0.06 | |
Hydra68 | Multi-CNN ensemble | 331.0 | TGRS2019 | 92.44 ± 0.34 | 94.51 ± 0.21 |
mmsCNN-HMM38 | 19.0 | RS2022 | 93.43 ± 0.25 | 95.51 ± 0.21 | |
MGML39 | None | TNNLS2023 | 90.69 ± 0.14 | 93.36 ± 0.12 | |
ESD-MBENet21 | 23.9 | TGRS2022 | 93.05 ± 0.18 | 95.36 ± 0.14 | |
RC-B3 | Single CNN | 12.2 | This Work | 94.69 ± 0.03 | 96.28 ± 0.05 |
Swin-T-Tiny | Single ViT | 28.3 | 94.41 ± 0.19 | 96.22 ± 0.13 | |
N-ViT-S | 31.7 | 94.81 ± 0.05 | 96.22 ± 0.07 | ||
RD-ESE | Dual-CNN ensemble | 17.5 | 95.17 ± 0.05 | 96.57 ± 0.06 | |
RST-ESE | CNN-ViT hybrid ensemble | 40.5 | 95.29 ± 0.04 | 96.78 ± 0.02 | |
RNV-ESE | 43.9 | 95.34 ± 0.08 | 96.70 ± 0.04 | ||
