SAR Target Classification on SAR Target Classification (20-shot)
88.1AccuracySG-MAE
Evaluation results
| Method | Links | |
|---|---|---|
| SG-MAEBackbone=Mamba, Params=27M, Pre-Training Setting=SSL-ImageNet & SAR, Pre-Training Dataset=ImageNet & SAR2026.06 | 88.1 | |
| SARATR-XBackbone=HiViT, Params=66M, Pre-Training Setting=SSL-ImageNet & SAR, Pre-Training Dataset=ImageNet & SAR2026.06 | 85.1 | |
| MAEBackbone=HiViT, Params=66M, Pre-Training Setting=SSL-ImageNet & SAR, Pre-Training Dataset=ImageNet & SAR2026.06 | 84 | |
| MAEBackbone=ViT, Params=86M, Pre-Training Setting=SSL-ImageNet & SAR, Pre-Training Dataset=ImageNet & SAR2026.06 | 83.6 | |
| MAEBackbone=Mamba, Params=27M, Pre-Training Setting=SSL-ImageNet & SAR, Pre-Training Dataset=ImageNet & SAR2026.06 | 82.1 | |
| MAEBackbone=HiViT, Params=66M, Pre-Training Setting=SSL-SAR, Pre-Training Dataset=SAR images2026.06 | 79.9 | |
| MAEBackbone=Mamba, Params=27M, Pre-Training Setting=SSL-SAR, Pre-Training Dataset=SAR images2026.06 | 76.7 | |
| SupervisedBackbone=ViT, Params=86M, Pre-Training Setting=SL-ImageNet, Pre-Training Dataset=ImageNet-1K2026.06 | 74.2 | |
| MAEBackbone=Mamba, Params=27M, Pre-Training Setting=SSL-ImageNet, Pre-Training Dataset=ImageNet-1K2026.06 | 70.88 | |
| SupervisedBackbone=ConvNeXt-V2, Params=89M, Pre-Training Setting=SL-ImageNet, Pre-Training Dataset=ImageNet-1K2026.06 | 70.5 | |
| FCMAEBackbone=ConvNeXt-V2, Params=89M, Pre-Training Setting=SSL-ImageNet & SAR, Pre-Training Dataset=ImageNet & SAR2026.06 | 69.5 | |
| MAEBackbone=HiViT, Params=66M, Pre-Training Setting=SSL-ImageNet, Pre-Training Dataset=ImageNet-1K2026.06 | 69.3 | |
| MAEBackbone=ViT, Params=86M, Pre-Training Setting=SSL-SAR, Pre-Training Dataset=SAR images2026.06 | 68.2 | |
| SupervisedBackbone=Mamba, Params=27M, Pre-Training Setting=SL-ImageNet, Pre-Training Dataset=ImageNet-1K2026.06 | 67.8 | |
| FCMAEBackbone=ConvNeXt-V2, Params=89M, Pre-Training Setting=SSL-SAR, Pre-Training Dataset=SAR images2026.06 | 67.7 | |
| MAEBackbone=ViT, Params=86M, Pre-Training Setting=SSL-ImageNet, Pre-Training Dataset=ImageNet-1K2026.06 | 65.5 | |
| FCMAEBackbone=ConvNeXt-V2, Params=89M, Pre-Training Setting=SSL-ImageNet, Pre-Training Dataset=ImageNet-1K2026.06 | 64 | |
| SupervisedBackbone=HiViT, Params=66M, Pre-Training Setting=SL-ImageNet, Pre-Training Dataset=ImageNet-1K2026.06 | 63.3 |