Open-Vocabulary Object Detection on OV-COCO
47.8AP@50 (Novel)CoT-PL
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CoT-PLBackbone=RN50×4 (87M), Supervision=Pseudo annotations2025.10 | 47.8 | 60.9 | |
| OV-DQUOBackbone=RN50×4 (87M), Supervision=CLIP Supervision2025.10 | 45.6 | — | |
| CLIP-SelfBackbone=ViT-L/14 (307M), Supervision=CLIP Supervision2025.10 | 44.3 | — | |
| CORA+Backbone=RN50×4 (87M), Supervision=CLIP Supervision2025.10 | 43.4 | 43.8 | |
| CoT-PLBackbone=RN50 (24M), Supervision=Pseudo annotations2025.10 | 43.4 | 58.9 | |
| CORA+Backbone=RN50×4 (87M), Supervision=COCO Caption2025.10 | 43.1 | 56.2 | |
| CCKT-DetBackbone=SwinB (88M), Supervision=CLIP Supervision2025.10 | 41.9 | 40.9 | |
| BINDBackbone=ViT-L/16 (307M), Supervision=CLIP Supervision2025.10 | 41.5 | 54.8 | |
| DeCo-DETRBackbone=RN50 (24M), Supervision=CLIP Supervision2025.10 | 41.3 | — | |
| LP-OVODBackbone=RN50 (24M), Supervision=Pseudo annotations2025.10 | 40.5 | 60.5 | |
| DITOBackbone=ViT-L/16(307M), Supervision=DataComp-1B2025.10 | 40.2 | 54.6 | |
| RegionCLIPBackbone=RN50×4 (87M), Supervision=CC3M2025.10 | 39.3 | 61.6 | |
| OV-DQUOBackbone=RN50 (24M), Supervision=CLIP Supervision2025.10 | 39.2 | — | |
| CAKEBackbone=RN50 (24M), Supervision=CLIP Supervision2025.10 | 38.2 | — | |
| CCKT-DetBackbone=RN50 (24M), Supervision=CLIP Supervision2025.10 | 38 | 35 | |
| LBPBackbone=RN50 (24M), Supervision=CLIP Supervision2025.10 | 37.8 | 58.7 | |
| CLIP-SelfBackbone=ViT-B/16 (86M), Supervision=CLIP Supervision2025.10 | 37.6 | — | |
| SAS-DetBackbone=RN50 (24M), Supervision=COCO Caption2025.10 | 37.4 | 58.5 | |
| OC-OVDBackbone=RN50 (24M), Supervision=COCO Caption2025.10 | 36.6 | 49.4 | |
| DITOBackbone=ViT-B/16(86M), Supervision=LAION-2B2025.10 | 36.6 | 48.8 | |
| BINDBackbone=ViT-B/16 (86M), Supervision=CLIP Supervision2025.10 | 36.3 | 50.2 | |
| RegionCLIPBackbone=RN50 (24M), Supervision=CC3M2025.10 | 35.2 | 57.6 | |
| CORABackbone=RN50 (24M), Supervision=CLIP Supervision2025.10 | 35.1 | 35.4 | |
| VL-PLMBackbone=RN50 (24M), Supervision=Pseudo annotations2025.10 | 34.4 | 60.2 | |
| CFM-ViTBackbone=ViT-L/16 (307M), Supervision=LAION-2B2025.10 | 34.3 | 46.4 | |
| BARONBackbone=RN50 (24M), Supervision=CLIP Supervision2025.10 | 34 | 60.4 | |
| PB-OVDBackbone=RN50 (24M), Supervision=COCO Caption2025.10 | 30.8 | 46.4 | |
| CoDetBackbone=RN50 (24M), Supervision=CC3M & COCO Caption2025.10 | 30.6 | 46.4 | |
| OV-DETRBackbone=RN50 (24M), Supervision=Pseudo annotations2025.10 | 29.4 | 52.7 | |
| DeticBackbone=RN50 (24M), Supervision=IN21K & CC3M2025.10 | 27.8 | 42 | |
| ViLD-ensBackbone=RN50 (24M), Supervision=CLIP Supervision2025.10 | 27.6 | 51.3 |