Semantic Segmentation on COCO 20^i (test)
67.9mIoUSegGPT
Evaluation Results
| Method | Links | |
|---|---|---|
| SegGPTshot=few-shot, model type=generalist2023.04 | 67.9 | |
| SegGPTshot=few-shot, category=generalist model, venue=ICCV'23, in-domain training=false, training data=includes COCO2023.05 | 67.9 | |
| FPTransshot=few-shot, model type=specialist, in-domain categories=true2023.04 | 65.5 | |
| Matchershot=few-shot, category=generalist model, venue=this work, in-domain training=false2023.05 | 60.7 | |
| FPTransshot=few-shot, model type=specialist, in-domain categories=false2023.04 | 58.9 | |
| FPTransshot=few-shot, category=specialist model, venue=NeurIPS'22, in-domain training=true2023.05 | 58.9 | |
| MSANetshot=few-shot, category=specialist model, venue=arXiv'22, in-domain training=true2023.05 | 56.8 | |
| FPTransshot=one-shot, model type=specialist, in-domain categories=true2023.04 | 56.5 | |
| SegGPTshot=one-shot, model type=generalist2023.04 | 56.1 | |
| SegGPTshot=1-shot, category=generalist model, venue=ICCV'23, in-domain training=false, training data=includes COCO2023.05 | 56.1 | |
| Matchershot=1-shot, category=generalist model, venue=this work, in-domain training=false2023.05 | 52.7 | |
| MSANetshot=1-shot, category=specialist model, venue=arXiv'22, in-domain training=true2023.05 | 51.1 | |
| HSNetshot=few-shot, model type=specialist, in-domain categories=true2023.04 | 50.7 | |
| FECANetShots=202023.01 | 50.3 | |
| FECANetShots=102023.01 | 49.6 | |
| HSNetshot=few-shot, model type=specialist, in-domain categories=false2023.04 | 49.5 | |
| HSNetshot=few-shot, category=specialist model, venue=ICCV'21, in-domain training=true2023.05 | 49.5 | |
| VATshot=few-shot, model type=specialist, in-domain categories=true2023.04 | 49.4 | |
| HSNetShots=102023.01 | 48.7 | |
| VATshot=few-shot, model type=specialist, in-domain categories=false2023.04 | 47.9 | |
| VATshot=few-shot, category=specialist model, venue=ECCV'22, in-domain training=true2023.05 | 47.9 | |
| FECANetShots=52023.01 | 47.6 | |
| FPTransshot=one-shot, model type=specialist, in-domain categories=false2023.04 | 47 | |
| FPTransshot=1-shot, category=specialist model, venue=NeurIPS'22, in-domain training=true2023.05 | 47 | |
| HSNetShots=52023.01 | 46.9 | |
| REPRIShots=102023.01 | 44.1 | |
| VATshot=one-shot, model type=specialist, in-domain categories=true2023.04 | 42.9 | |
| HSNetshot=one-shot, model type=specialist, in-domain categories=true2023.04 | 41.7 | |
| REPRIShots=52023.01 | 41.6 | |
| FECANetShots=12023.01 | 41.6 | |
| VATshot=one-shot, model type=specialist, in-domain categories=false2023.04 | 41.3 | |
| VATshot=1-shot, category=specialist model, venue=ECCV'22, in-domain training=true2023.05 | 41.3 | |
| HSNetshot=one-shot, model type=specialist, in-domain categories=false2023.04 | 41.2 | |
| HSNetshot=1-shot, category=specialist model, venue=ICCV'21, in-domain training=true2023.05 | 41.2 | |
| PFENetShots=102023.01 | 39.7 | |
| HSNetShots=12023.01 | 39.2 | |
| PFENetShots=52023.01 | 39 | |
| PFENetShots=12023.01 | 35.8 | |
| RPMMShots=52023.01 | 35.5 | |
| REPRIShots=12023.01 | 34.1 | |
| RPMMShots=102023.01 | 33.1 | |
| Paintershot=1-shot, category=generalist model, venue=CVPR'23, in-domain training=false2023.05 | 33.1 | |
| Paintershot=one-shot, model type=generalist2023.04 | 32.8 | |
| Paintershot=few-shot, model type=generalist2023.04 | 32.6 | |
| Paintershot=few-shot, category=generalist model, venue=CVPR'23, in-domain training=false2023.05 | 32.6 | |
| RPMMShots=12023.01 | 30.6 | |
| PerSAM-Fshot=1-shot, category=generalist model, venue=arXiv'23, in-domain training=false, training-free=true2023.05 | 23.5 | |
| PerSAMshot=1-shot, category=generalist model, venue=arXiv'23, in-domain training=false, using SAM=true2023.05 | 23 |