Semantic Correspondence on AP-10K C.S.
88.3PCK@0.10MARCO
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MARCOSupervision=Supervised, DINOv2 Encoder=true, SD Encoder=false, Uses object masks at training=true2026.04 | 88.3 | 32.2 | 76.6 | |
| MARCOSupervision=Supervised, DINOv2 Encoder=true, SD Encoder=false2026.04 | 87.5 | 31.3 | 74.7 | |
| Geo-SCSupervision=Supervised, DINOv2 Encoder=true, SD Encoder=true, Uses object masks at inference=true2026.04 | 85.9 | 21.7 | 70.3 | |
| GECOSupervision=Supervised, DINOv2 Encoder=false, SD Encoder=true, Uses object masks at training=true2026.04 | 81.2 | 17.4 | 64.9 | |
| SD+DINOv2Supervision=Supervised, DINOv2 Encoder=true, SD Encoder=true2026.04 | 74 | 8.8 | 53.9 | |
| DIY-SCSupervision=Unsupervised / Weakly Supervised, DINOv2 Encoder=true, SD Encoder=true, Uses object masks at training=true2026.04 | 69.8 | — | — | |
| DHFSupervision=Supervised, DINOv2 Encoder=false, SD Encoder=true2026.04 | 60 | 6.8 | 42.4 | |
| SD+DINOv2Supervision=Unsupervised / Weakly Supervised, DINOv2 Encoder=true, SD Encoder=true2026.04 | 59.3 | 6.4 | 39.7 | |
| DINOv2+NNSupervision=Unsupervised / Weakly Supervised, DINOv2 Encoder=true, SD Encoder=false2026.04 | 57.3 | 5.3 | 37 | |
| DIFTSupervision=Unsupervised / Weakly Supervised, DINOv2 Encoder=false, SD Encoder=true2026.04 | 46 | 5.1 | 30.8 |