Semantic Correspondence on SPair-71k (PCK variants)
76.86Φ_bbox @ α=0.1Ours-XL-t
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| Ours-XL-tCategory=Ours, Backbone=SDXL, Enhancements=Attention maps [58], Feature amalgamation [2, 29, 60]2024.10 | 76.86 | — | — | 83.9 | — | — | — | |
| SD4Match-CPMTraining/Tuning Dataset=SPair-71k2023.10 | 75.5 | — | — | — | — | 59.5 | — | |
| Ours-XLCategory=Ours, Backbone=SDXL2024.10 | 75.18 | — | — | 81.72 | — | — | — | |
| SD4Match-SingleTraining/Tuning Dataset=SPair-71k2023.10 | 72.6 | — | — | — | — | 56.5 | — | |
| Ours-v1.5Category=Ours, Backbone=SDv1.52024.10 | 69.83 | — | — | 77.78 | — | — | — | |
| Legacy-v1.5Category=Baseline, Backbone=SDv1.52024.10 | 66.73 | — | — | 75.14 | — | — | — | |
| DHFCategory=SOTA2024.10 | 64.61 | — | — | 72.56 | — | — | — | |
| GeoAware-SCBackbone=SD+DINOv2, Reso.=960, 840, Evaluation Protocol=zero-shot2025.05 | 61.3 | — | — | — | — | 45.3 | — | |
| OursBackbone=SD+DINOv2, Reso.=960, 840, Evaluation Protocol=zero-shot2025.05 | 60.4 | — | — | — | — | 47.9 | 67 | |
| SD+DINOTraining/Tuning Dataset=PF-Pascal (default/zero-shot)2023.10 | 59.3 | — | — | — | — | 44.1 | — | |
| SD-DINOBackbone=SD+DINOv2, Reso.=960, 840, Evaluation Protocol=zero-shot2025.05 | 59.3 | — | — | — | — | — | — | |
| Legacy-XLCategory=Baseline, Backbone=SDXL2024.10 | 59.16 | — | — | 66 | — | — | — | |
| DINOv2Training/Tuning Dataset=PF-Pascal (default/zero-shot)2023.10 | 53.9 | — | — | — | — | 38.4 | — | |
| DIFTTraining/Tuning Dataset=PF-Pascal (default/zero-shot)2023.10 | 52.9 | — | — | — | — | 39.7 | — | |
| DIFTCategory=SOTA2024.10 | 52.9 | — | — | — | — | — | — | |
| DIFTsdBackbone=SD, Reso.=768, Evaluation Protocol=zero-shot2025.05 | 52.9 | — | — | — | — | — | — | |
| LDMBackbone=SD, Evaluation Protocol=zero-shot2025.05 | 45.4 | — | — | — | — | 28.9 | — | |
| SD4Match-SingleTraining/Tuning Dataset=PF-Pascal2023.10 | 44.4 | — | — | — | — | 30.5 | — | |
| DHPFCategory=SOTA2024.10 | 42.63 | — | — | 55.28 | — | — | — | |
| DINOCategory=SOTA2024.10 | 41.04 | — | — | 51.68 | — | — | — | |
| SD4Match-CPMTraining/Tuning Dataset=PF-Pascal2023.10 | 40.9 | — | — | — | — | 27.2 | — | |
| SimSC-iBOTTraining/Tuning Dataset=PF-Pascal2023.10 | 37.9 | — | — | — | — | 22 | — | |
| DINOv1Training/Tuning Dataset=PF-Pascal (default/zero-shot)2023.10 | 35.9 | — | — | — | — | 23.1 | — | |
| PWarpC-NCNetTraining/Tuning Dataset=PF-Pascal2023.10 | 35.3 | — | — | — | — | 18.2 | — | |
| Trans.MatTraining/Tuning Dataset=PF-Pascal2023.10 | 30.1 | — | — | — | — | — | — | |
| DHPFTraining/Tuning Dataset=PF-Pascal2023.10 | 27.5 | — | — | — | — | 15.3 | — | |
| CATSTraining/Tuning Dataset=PF-Pascal2023.10 | 27 | — | — | — | — | 13.6 | — | |
| NCNetTraining/Tuning Dataset=PF-Pascal2023.10 | 26.4 | — | — | — | — | — | — | |
| SFNetTraining/Tuning Dataset=PF-Pascal2023.10 | 24 | — | — | — | — | 11.2 | — | |
| CATSF.T.=true, Feat.=true, Data Aug.=false, Cost Aggregation=Transformer2022.07 | — | 10.2 | 21.6 | 43.5 | 55 | — | — | |
| CATSF.T.=true, Feat.=true, Data Aug.=true, Cost Aggregation=Transformer2022.07 | — | 13.8 | 27.7 | 49.9 | 61.7 | — | — | |
| CHMF.T.=false, Feat.=true, Data Aug.=false, Cost Aggregation=4D Conv.2022.07 | — | 14.9 | 27.2 | 46.3 | 57.5 | — | — | |
| DHPFF.T.=false, Feat.=true, Data Aug.=false, Cost Aggregation=RHM2022.07 | — | 11 | 20.9 | 37.3 | 47.5 | — | — | |
| DHPFF.T.=false, Feat.=true, Data Aug.=true, Cost Aggregation=RHM2022.07 | — | — | — | 39.4 | — | — | — | |
| MMNetF.T.=true, Feat.=true, Data Aug.=false2022.07 | — | — | — | 40.9 | — | — | — | |
| NC-NetF.T.=true, Feat.=true, Data Aug.=false, Cost Aggregation=4D Conv.2022.07 | — | — | — | — | 20.1 | — | — | |
| PMNCF.T.=true, Feat.=true, Data Aug.=false, Cost Aggregation=4D Conv.2022.07 | — | — | — | 50.4 | — | — | — | |
| SCOTF.T.=false, Feat.=true, Data Aug.=false, Cost Aggregation=OT-RHM2022.07 | — | — | — | 35.6 | — | — | — | |
| VATF.T.=true, Feat.=true, Data Aug.=false, Cost Aggregation=Transformer2022.07 | — | 14.9 | 28.3 | 48.4 | 59.1 | — | — | |
| VATF.T.=true, Feat.=true, Data Aug.=true, Cost Aggregation=Transformer2022.07 | — | 19.6 | 35 | 55.5 | 65.1 | — | — |