Semantic Segmentation on iSAID 5i
52.4mIoUGF-SAM
Evaluation results
| Method | Links | |
|---|---|---|
| GF-SAMModel Category=generalist model, Shots=5-shot2024.10 | 52.4 | |
| INSID3Encoder=DINOv3, #Param=304 M, Training Protocol=Training free, Supervision Type=Unsupervised pre-training2026.03 | 52.1 | |
| GF-SAM† + debiasEncoder=DINOv3 + SAM, #Param=945 M, Training Protocol=Training free, Supervision Type=Mask-supervised pre-training, Debias=True2026.03 | 47.6 | |
| DiffewSEncoder=Stable Diffusion, #Param=890 M, Training Protocol=Task-specific fine-tuning, Supervision Type=Semantic + mask supervision2026.03 | 47.5 | |
| GF-SAMModel Category=generalist model, Shots=1-shot2024.10 | 47.1 | |
| GF-SAMEncoder=DINOv2 + SAM, #Param=945 M, Training Protocol=Training free, Supervision Type=Mask-supervised pre-training2026.03 | 47.1 | |
| GF-SAM†Encoder=DINOv3 + SAM, #Param=945 M, Training Protocol=Training free, Supervision Type=Mask-supervised pre-training2026.03 | 46.7 | |
| SegICEncoder=DINOv2, #Param=310 M, Training Protocol=Task-specific fine-tuning, Supervision Type=Semantic + mask supervision2026.03 | 46.1 | |
| FRINetModel Category=specialist model, Shots=5-shot2024.10 | 44.5 | |
| FRINetModel Category=specialist model, Shots=1-shot2024.10 | 42.6 | |
| SegIC (COCO)Encoder=DINOv2, #Param=310 M, Training Protocol=Task-specific fine-tuning, Supervision Type=Semantic + mask supervision, dataset split=COCO2026.03 | 40.8 | |
| HSNetModel Category=specialist model, Shots=5-shot2024.10 | 40.4 | |
| SINEEncoder=DINOv2, #Param=373 M, Training Protocol=Task-specific fine-tuning, Supervision Type=Semantic + mask supervision2026.03 | 38.3 | |
| MatcherModel Category=generalist model, Shots=5-shot2024.10 | 34.3 | |
| HSNetModel Category=specialist model, Shots=1-shot2024.10 | 34.1 | |
| MatcherModel Category=generalist model, Shots=1-shot2024.10 | 33.3 | |
| MatcherEncoder=DINOv2 + SAM, #Param=945 M, Training Protocol=Training free, Supervision Type=Mask-supervised pre-training2026.03 | 33.3 | |
| SegGPTEncoder=ViT, #Param=354 M, Training Protocol=Task-specific fine-tuning, Supervision Type=Semantic + mask supervision2026.03 | 30.9 | |
| PerSAM-FModel Category=generalist model, Shots=1-shot2024.10 | 20.3 | |
| PerSAMModel Category=generalist model, Shots=1-shot2024.10 | 19.2 | |
| PerSAMEncoder=SAM, #Param=640 M, Training Protocol=Training free, Supervision Type=Mask-supervised pre-training2026.03 | 19.2 |