Scene Classification on HRSSC (test)
85.22OASpecAware
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| SpecAwarePretraining dataset=Hyper-400K, Evaluation protocol=fine-tuning2025.10 | 85.22 | 82.86 | 75.01 | |
| SpectralEarthPretraining dataset=SpectralEarth, Evaluation protocol=fine-tuning2025.10 | 84.61 | 82.07 | 71.01 | |
| SpecAwarePretraining dataset=Hyper-400K, Evaluation protocol=linear probing2025.10 | 84 | 81.34 | 71.55 | |
| HyperSIGMAPretraining dataset=HyperGlobal-450K2025.10 | 81.85 | 78.88 | 68.45 | |
| SpectralEarthPretraining dataset=SpectralEarth, Evaluation protocol=linear probing2025.10 | 80.78 | 77.52 | 69.57 | |
| SatMAEPretraining dataset=fMoW-S22025.10 | 80.32 | 77.13 | 66.55 | |
| SpectralGPT+Pretraining dataset=fMoW-S2+BigEarthNet2025.10 | 79.86 | 76.58 | 66.55 | |
| SatMAE++Pretraining dataset=fMoW-S22025.10 | 79.1 | 75.67 | 66.07 | |
| ViT-BasePretraining dataset=ImageNet-1K2025.10 | 78.71 | 75.16 | 66.43 | |
| SpecResNet50Pretraining dataset=SpectralEarth2025.10 | 66.16 | 59.62 | 41.8 | |
| ResNet101Pretraining dataset=ImageNet-1K2025.10 | 57.2 | 50.52 | 46.89 |