Building Extraction on Inria
78.61mIoUPi-Seg
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Pi-SegBackbone=ViT-B, Type=OVRSIS2026.04 | 78.61 | 90.55 | |
| Pi-SegBackbone=ViT-L, Type=OVRSIS2026.04 | 77.82 | 91.58 | |
| OVRSBackbone=ViT-L, Type=OVRSIS2026.04 | 76.17 | 89.96 | |
| CAT-SEGBackbone=ViT-B, Type=OVS2026.04 | 75.87 | 90.29 | |
| OVRSBackbone=ViT-B, Type=OVRSIS2026.04 | 75.73 | 88.02 | |
| CAT-SEGBackbone=ViT-L, Type=OVS2026.04 | 75.11 | 91.15 | |
| SANBackbone=ViT-L, Type=OVS2026.04 | 75.03 | 90.06 | |
| GSNETBackbone=ViT-B, Type=OVRSIS2026.04 | 73.78 | 90.06 | |
| FCCLIPBackbone=ConvNeXt-B, Type=OVS2026.04 | 73.02 | 89.59 | |
| GSNETBackbone=ViT-L, Type=OVRSIS2026.04 | 72.12 | 90.69 | |
| RSKT-SEGBackbone=ViT-L, Type=OVRSIS2026.04 | 72.12 | 90.69 | |
| SCANBackbone=ViT-L, Type=OVS2026.04 | 70.87 | 82.25 | |
| RSKT-SEGBackbone=ViT-B, Type=OVRSIS2026.04 | 70.68 | 88.57 | |
| EOV-SEGBackbone=ConvNeXt-B, Type=OVS2026.04 | 68.32 | 89.57 | |
| SANBackbone=ViT-B, Type=OVS2026.04 | 63.64 | 84.49 | |
| SEDBackbone=ConvNeXt-B, Type=OVS2026.04 | 61.61 | 86.46 | |
| FCCLIPBackbone=ConvNeXt-L, Type=OVS2026.04 | 61.23 | 86.04 | |
| SCANBackbone=ViT-B, Type=OVS2026.04 | 60.19 | 81.31 | |
| SEDBackbone=ConvNeXt-L, Type=OVS2026.04 | 58.77 | 84.89 | |
| EOV-SEGBackbone=ConvNeXt-L, Type=OVS2026.04 | 56.47 | 81.29 | |
| ConInfer2026.03 | 55.65 | — | |
| CLIP-SAMTraining Protocol=Zero-shot, Input Resolution=896x8962026.02 | 53.8 | — | |
| SegEarth-OVResolution=896 x 8962024.10 | 48.9 | — | |
| SegEarth-OVTraining Protocol=Trained on remote sensing data, Input Resolution=896x8962026.02 | 48.9 | — | |
| OursTraining/Evaluation Protocol=Zero-shot2026.02 | 48 | — | |
| SegEarth-OV2026.03 | 45.41 | — | |
| SegEarth-OVResolution=448 x 4482024.10 | 44.6 | — | |
| SegEarth-OVTraining/Evaluation Protocol=Trained on remote sensing data2026.02 | 44.6 | — | |
| SegEarth-OVBackbone=ViT-B, Type=OVRSIS (TF)2026.04 | 44.6 | — | |
| GEM2026.03 | 40.74 | — | |
| ClearCLIPResolution=448 x 4482024.10 | 39 | — | |
| ClearCLIPTraining/Evaluation Protocol=Zero-shot2026.02 | 39 | — | |
| ClearCLIPBackbone=ViT-B, Type=OVS (TF)2026.04 | 39 | — | |
| MaskCLIP*configuration=baseline2026.03 | 37.26 | — | |
| ClearCLIP2026.03 | 37.08 | — | |
| SCLIPResolution=448 x 4482024.10 | 34.9 | — | |
| SCLIPTraining/Evaluation Protocol=Zero-shot2026.02 | 34.9 | — | |
| SCLIPBackbone=ViT-B, Type=OVS (TF)2026.04 | 34.9 | — | |
| SCLIP2026.03 | 33.92 | — | |
| MaskCLIPResolution=448 x 4482024.10 | 33.4 | — | |
| MaskCLIPTraining/Evaluation Protocol=Zero-shot2026.02 | 33.4 | — | |
| MaskCLIPBackbone=ViT-B, Type=OVS (TF)2026.04 | 33.4 | — | |
| MaskCLIP2026.03 | 33.15 | — | |
| GEMResolution=448 x 4482024.10 | 28.5 | — | |
| GEMTraining/Evaluation Protocol=Zero-shot2026.02 | 28.5 | — | |
| GEMBackbone=ViT-B, Type=OVS (TF)2026.04 | 28.5 | — | |
| CLIPResolution=448 x 4482024.10 | 19.6 | — | |
| CLIPTraining/Evaluation Protocol=Zero-shot2026.02 | 19.6 | — | |
| CLIPBackbone=ViT-B, Type=OVS (TF)2026.04 | 19.6 | — | |
| CLIP2026.03 | 15.12 | — |