Open-vocabulary Object Detection on COCO (val)
67.23AP50 (All)OvarNet
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| OvarNetEvaluation Setting=box-given, Training Data=+COCO Cap-sub2023.01 | 67.23 | 71.88 | 54.1 | |
| OvarNetEvaluation Setting=box-given, Training Data=+CC 3M-sub2023.01 | 64.18 | 68.35 | 52.34 | |
| OvarNetEvaluation Setting=box-given, Training Data=COCO-base + VAW-base2023.01 | 55.85 | 60.94 | 41.44 | |
| OvarNetEvaluation Setting=box-free, Training Data=+COCO Cap-sub2023.01 | 54.15 | 60.35 | 35.17 | |
| OvarNetEvaluation Setting=box-free, Training Data=+CC 3M-sub2023.01 | 52.4 | 59.5 | 33.68 | |
| ViLDTraining Data=CLIP400M2023.01 | 51.3 | 59.5 | 27.6 | |
| PromptDetTraining Data=Web Images2023.01 | 50.6 | — | 26.6 | |
| Region CLIPTraining Data=CC 3M2023.01 | 50.4 | 57.1 | 31.4 | |
| OvarNetEvaluation Setting=box-free, Training Data=COCO-base + VAW-base2023.01 | 49.77 | 56.2 | 32.02 | |
| Region CLIPTraining Data=COCO Cap2023.01 | 47.5 | 54.8 | 26.8 | |
| DeticTraining Data=COCO Cap2023.01 | 45 | 47.1 | 27.8 | |
| OVR-RCNNTraining Data=CC 3M2023.01 | 39.9 | 46 | 22.8 | |
| OVR-RCNNTraining Data=COCO Cap2023.01 | 34.3 | — | — |