Object Detection on DX-20
11.5LaECE0FRCNN-IR
Evaluation results
| Method | Links | ||
|---|---|---|---|
| FRCNN-IRDetector Architecture=Faster R-CNN, Calibration=Isotonic Regression2026.05 | 11.5 | 27 | |
| YOLOX-PSDetector Architecture=YOLOX, Calibration=Platt Scaling2026.05 | 13.8 | 18.8 | |
| FRCNN-PSDetector Architecture=Faster R-CNN, Calibration=Platt Scaling2026.05 | 14.2 | 27.3 | |
| YOLOX-IRDetector Architecture=YOLOX, Calibration=Isotonic Regression2026.05 | 14.4 | 19.1 | |
| YOLOX-DEDetector Architecture=YOLOX, Calibration=Deep Ensembles2026.05 | 14.5 | 20.8 | |
| YOLOX-G†Detector Architecture=YOLOX, Calibration=Proposed G†2026.05 | 14.5 | 21.2 | |
| RetinaNet-IRDetector Architecture=RetinaNet, Calibration=Isotonic Regression2026.05 | 14.7 | 27.1 | |
| RetinaNet-G†Detector Architecture=RetinaNet, Calibration=Proposed G†2026.05 | 15.3 | 29.7 | |
| RetinaNet-DEDetector Architecture=RetinaNet, Calibration=Deep Ensembles2026.05 | 15.5 | 28.9 | |
| RetinaNet-PSDetector Architecture=RetinaNet, Calibration=Platt Scaling2026.05 | 15.9 | 27.4 | |
| FRCNN-DEDetector Architecture=Faster R-CNN, Calibration=Deep Ensembles2026.05 | 16.3 | 27.7 | |
| FRCNN-G†Detector Architecture=Faster R-CNN, Calibration=Proposed G†2026.05 | 16.7 | 27.6 | |
| YOLOXDetector Architecture=YOLOX, Calibration=None2026.05 | 18.5 | 27.8 | |
| FRCNNDetector Architecture=Faster R-CNN, Calibration=None2026.05 | 20.7 | 30 | |
| RetinaNetDetector Architecture=RetinaNet, Calibration=None2026.05 | 26.8 | 40.7 |