Object Detection on VIPER (val)
31.8mAP50Dataset-specific (Oracle)
Evaluation results
| Method | Links | |
|---|---|---|
| Dataset-specific (Oracle)Training Data=VIPER training set, Backbone=ResNet-502021.02 | 31.8 | |
| Unified detector (UniDet)Strategy=Unified (retrained), Backbone=ResNet-502021.02 | 21.3 | |
| Partitioned DetectorStrategy=Partitioned, Backbone=ResNet-502021.02 | 20.9 | |
| Objects365 modelTraining Data=Objects365, Backbone=ResNet-502021.02 | 20.7 | |
| EnsembleStrategy=Ensemble, Backbone=ResNet-502021.02 | 16.8 | |
| Mapillary modelTraining Data=Mapillary, Backbone=ResNet-502021.02 | 15.2 | |
| COCO modelTraining Data=COCO, Backbone=ResNet-502021.02 | 13.9 | |
| OpenImages modelTraining Data=OpenImages, Backbone=ResNet-502021.02 | 10.4 |