Object Detection on SpaceDet Domain 1
50.09PrecisionEWC
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| EWCParadigm=Continual Learning2026.03 | 50.09 | 32.62 | 39.51 | 21.31 | |
| Single Domain 1Training Paradigm=Single-domain training (Domain 1)2026.03 | 49.38 | 47 | 48.16 | 47.86 | |
| Joint TrainingTraining Paradigm=Joint multi-domain training2026.03 | 48.42 | 48 | 48.21 | 47.05 | |
| Dual-Stage Invariant Continual LearningMethod Label=Ours2026.03 | 44.99 | 47 | 45.97 | 44.22 | |
| Shmelkov et al.Paradigm=Continual Learning2026.03 | 42.57 | 45 | 43.75 | 37.84 | |
| CWDParadigm=Continual Learning2026.03 | 42.25 | 27.06 | 32.99 | 22.31 | |
| OGDParadigm=Continual Learning2026.03 | 41.81 | 29.23 | 34.41 | 19.09 | |
| YOLOv10mDetector=YOLOv10m2026.03 | 40.39 | 38 | 39.15 | 37.71 | |
| Deformable DETRDetector=Deformable DETR2026.03 | 38.62 | 39 | 38.81 | 36.89 | |
| Sequential w/o CLTraining Paradigm=Sequential fine-tuning2026.03 | 37.92 | 37 | 37.46 | 34.28 | |
| Faster R-CNNDetector=Faster R-CNN2026.03 | 37.92 | 37 | 37.46 | 34.28 | |
| Single Domain 2Training Paradigm=Single-domain training (Domain 2)2026.03 | 31.69 | 33 | 32.33 | 28.98 | |
| CL-DETRParadigm=Continual Learning2026.03 | 26.72 | 13 | 17.49 | 9.28 | |
| DETRDetector=DETR2026.03 | 6.54 | 10 | 7.91 | 1.59 |