Object Detection on SpaceDet Total Metric
63.27PrecisionEWC
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| EWCParadigm=Continual Learning2026.03 | 63.27 | 34.71 | 44.59 | 33.99 | |
| OGDParadigm=Continual Learning2026.03 | 62.65 | 31.69 | 41.43 | 31.85 | |
| CWDParadigm=Continual Learning2026.03 | 60.38 | 28.73 | 38.41 | 33.24 | |
| Joint TrainingTraining Paradigm=Joint multi-domain training2026.03 | 50.24 | 46.5 | 48.24 | 42.56 | |
| Dual-Stage Invariant Continual LearningMethod Label=Ours2026.03 | 49.36 | 47.5 | 48.34 | 42.62 | |
| Shmelkov et al.Paradigm=Continual Learning2026.03 | 47.56 | 44.5 | 45.82 | 37.31 | |
| Deformable DETRDetector=Deformable DETR2026.03 | 46.24 | 41.5 | 43.62 | 38.6 | |
| Single Domain 1Training Paradigm=Single-domain training (Domain 1)2026.03 | 45.61 | 44 | 44.79 | 39.69 | |
| YOLOv10mDetector=YOLOv10m2026.03 | 44.47 | 42 | 43.19 | 37.14 | |
| Sequential w/o CLTraining Paradigm=Sequential fine-tuning2026.03 | 41.09 | 42 | 41.52 | 35.23 | |
| Faster R-CNNDetector=Faster R-CNN2026.03 | 41.09 | 42 | 41.52 | 35.23 | |
| Single Domain 2Training Paradigm=Single-domain training (Domain 2)2026.03 | 40.34 | 38.5 | 39.35 | 32.65 | |
| CL-DETRParadigm=Continual Learning2026.03 | 25.61 | 18.5 | 20.87 | 8.21 | |
| DETRDetector=DETR2026.03 | 7.87 | 13 | 9.8 | 1.81 |