Semantic Segmentation on UDD5
71.7mIoUSegEarth-OV3
Evaluation results
| Method | Links | ||
|---|---|---|---|
| SegEarth-OV3Training Regime=Training-free2025.12 | 71.7 | — | |
| Pi-SegBackbone=ViT-L, Type=OVRSIS, Training Dataset=DLRSD2026.04 | 59.08 | 72.58 | |
| OracleTraining Regime=Oracle (fully supervised SegFormer-b0)2025.12 | 56.5 | — | |
| Pi-SegBackbone=ViT-B, Type=OVRSIS, Training Dataset=DLRSD2026.04 | 52.47 | 67.3 | |
| SegEarth-OVVenue=Ours2024.10 | 50.6 | — | |
| SegEarth-OVTraining Regime=Training-free2025.12 | 50.6 | — | |
| SegEarth-OV2026.03 | 49.64 | — | |
| ConInfer2026.03 | 46.86 | — | |
| DR-SegBackbone=ViT-L, Training Dataset=DLRSD2026.04 | 46.63 | 62.61 | |
| DR-SegBackbone=ViT-B, Training Dataset=DLRSD2026.04 | 46.36 | 62.14 | |
| CorrCLIPTraining Regime=Training-free2025.12 | 46.1 | — | |
| RSKT-SegBackbone=ViT-L, Training Dataset=DLRSD2026.04 | 45.73 | 59.4 | |
| RSKT-SegBackbone=ViT-B, Training Dataset=DLRSD2026.04 | 42.99 | 57.81 | |
| MaskCLIP*type=baseline2026.03 | 42.17 | — | |
| RSKT-SegTraining Regime=Training on remote sensing segmentation data2025.12 | 42.1 | — | |
| ClearCLIPVenue=ECCV'242024.10 | 41.8 | — | |
| ClearCLIPTraining Regime=Training-free2025.12 | 41.8 | — | |
| GEMVenue=CVPR 242024.10 | 41.2 | — | |
| GEMTraining Regime=Training-free2025.12 | 41.2 | — | |
| GSNetBackbone=ViT-L, Training Dataset=DLRSD2026.04 | 40.92 | 57.04 | |
| GSNetTraining Regime=Training on remote sensing segmentation data2025.12 | 40.9 | — | |
| OVRSBackbone=ViT-L, Training Dataset=DLRSD2026.04 | 40.82 | 54.24 | |
| OVRSTraining Regime=Training on remote sensing segmentation data2025.12 | 40.8 | — | |
| ProxyCLIPTraining Regime=Training-free2025.12 | 40.8 | — | |
| Cat-SegBackbone=ViT-L, Training Dataset=DLRSD2026.04 | 40.24 | 51.65 | |
| Cat-SegTraining Regime=Training on remote sensing segmentation data2025.12 | 40.2 | — | |
| GSNetBackbone=ViT-B, Training Dataset=DLRSD2026.04 | 40.05 | 51.72 | |
| OVRSBackbone=ViT-B, Training Dataset=DLRSD2026.04 | 39.1 | 50.65 | |
| SCLIPVenue=arXiv'232024.10 | 38.7 | — | |
| SCLIPTraining Regime=Training-free2025.12 | 38.7 | — | |
| Cat-SegBackbone=ViT-B, Training Dataset=DLRSD2026.04 | 38.23 | 49.56 | |
| ClearCLIP2026.03 | 38.07 | — | |
| Pi-SegBackbone=ViT-L, Type=OVRSIS, Training Dataset=iSAID2026.04 | 37.24 | 53.87 | |
| SANBackbone=ViT-L, Training Dataset=DLRSD2026.04 | 37.23 | 48.45 | |
| SANTraining Regime=Training on remote sensing segmentation data2025.12 | 37.2 | — | |
| SANBackbone=ViT-B, Training Dataset=DLRSD2026.04 | 36.87 | 47.21 | |
| GEM2026.03 | 36.02 | — | |
| SEDBackbone=ConvNeXt-L, Training Dataset=DLRSD2026.04 | 35.73 | 45.15 | |
| SEDTraining Regime=Training on remote sensing segmentation data2025.12 | 35.7 | — | |
| SEDBackbone=ConvNeXt-B, Training Dataset=DLRSD2026.04 | 34.65 | 44.1 | |
| SCANBackbone=ViT-L, Training Dataset=DLRSD2026.04 | 34.14 | 43.25 | |
| SCANTraining Regime=Training on remote sensing segmentation data2025.12 | 34.1 | — | |
| RSKT-SegBackbone=ViT-L, Training Dataset=iSAID2026.04 | 34.01 | 57.38 | |
| DR-SegBackbone=ViT-L, Training Dataset=iSAID2026.04 | 33.42 | 59.82 | |
| SCLIP2026.03 | 33.25 | — | |
| MaskCLIP2026.03 | 32.74 | — | |
| MaskCLIPVenue=ECCV'222024.10 | 32.4 | — | |
| MaskCLIPTraining Regime=Training-free2025.12 | 32.4 | — | |
| GSNetBackbone=ViT-L, Training Dataset=iSAID2026.04 | 32.24 | 56.2 | |
| SCANBackbone=ViT-B, Training Dataset=DLRSD2026.04 | 32.12 | 40.45 | |
| OVRSBackbone=ViT-L, Training Dataset=iSAID2026.04 | 31.88 | 55.32 | |
| DR-SegBackbone=ViT-B, Training Dataset=iSAID2026.04 | 31.12 | 58.44 | |
| Cat-SegBackbone=ViT-L, Training Dataset=iSAID2026.04 | 24.32 | 42.77 | |
| SEDBackbone=ConvNeXt-L, Training Dataset=iSAID2026.04 | 22.42 | 39.15 | |
| SANBackbone=ViT-L, Training Dataset=iSAID2026.04 | 21.35 | 37.66 | |
| SCANBackbone=ViT-L, Training Dataset=iSAID2026.04 | 16.22 | 32.51 | |
| RSKT-SegBackbone=ViT-B, Training Dataset=iSAID2026.04 | 13.94 | 43.47 | |
| GSNetBackbone=ViT-B, Training Dataset=iSAID2026.04 | 12.44 | 39.62 | |
| SANBackbone=ViT-B, Training Dataset=iSAID2026.04 | 12.11 | 32.45 | |
| Cat-SegBackbone=ViT-B, Training Dataset=iSAID2026.04 | 12.1 | 38.55 | |
| OVRSBackbone=ViT-B, Training Dataset=iSAID2026.04 | 11.9 | 38.8 | |
| SEDBackbone=ConvNeXt-B, Training Dataset=iSAID2026.04 | 10.23 | 31.5 | |
| CLIPVenue=ICML 212024.10 | 9.5 | — | |
| CLIPTraining Regime=Training-free2025.12 | 9.5 | — | |
| SCANBackbone=ViT-B, Training Dataset=iSAID2026.04 | 8.45 | 28.6 | |
| CLIP2026.03 | 7.91 | — |