Open World Object Detection on OWOD Evaluation Protocol (VOC + MS-COCO) Task 1 2007 (test)
6.03WIFR-CNN
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| FR-CNNuse unknown's annotation in train set=false, use unknown's annotation in val set=false, Backbone=ResNet-50 FPN2022.03 | 6.03 | 8,468 | — | 58.81 | |
| OREuse unknown's annotation in train set=false, use unknown's annotation in val set=true, Backbone=ResNet-50 FPN2022.03 | 5.11 | 6,833 | — | 58.93 | |
| OpenDetuse unknown's annotation in train set=false, use unknown's annotation in val set=false, Backbone=ResNet-50 FPN2022.03 | 4.44 | 5,781 | — | 59.01 | |
| FR-CNN (Oracle)use unknown's annotation in train set=false, use unknown's annotation in val set=true, Backbone=ResNet-50 FPN2022.03 | 4.27 | 6,862 | — | 60.43 | |
| Faster-RCNN2021.03 | 0.0699 | 13,396 | 56.16 | — | |
| ORE2021.03 | 0.0219 | 8,234 | 56.34 | — | |
| Oracle2021.03 | 0.02 | 7,080 | 57.76 | — |