Object Pose Estimation on T-LESS (seen (obj. 1-18) and novel (obj. 19-30))
70.87VSD Recall (Seen)context-aware weight aggregation network
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| context-aware weight aggregation networkNumber templates=21 K, Aggregation Mode=CWA2024.03 | 70.87 | 71.26 | 71.03 | |
| Nguyen et al.Reference Source=GT CAD, Ref. image sampling=None2023.03 | 60.15 | 58.7 | 59.57 | |
| Template-poseNumber templates=92 K2024.03 | 59.62 | 57.75 | 58.87 | |
| Template-poseNumber templates=21 K2024.03 | 59.14 | 56.91 | 58.25 | |
| NOPEReference Source=1 ref. image, Ref. image sampling=front views2023.03 | 49.3 | 48.46 | 48.96 | |
| NOPEReference Source=1 ref. image, Ref. image sampling=all views2023.03 | 47.03 | 45.69 | 46.49 | |
| MultiPathReference Source=GT CAD, Ref. image sampling=None2023.03 | 43.17 | 43.33 | 43.24 | |
| ImplicitNumber templates=92 K2024.03 | 35.6 | 42.45 | 38.34 | |
| MPLNumber templates=92 K2024.03 | 35.25 | 33.17 | 34.42 | |
| PIZZAReference Source=1 ref. image, Ref. image sampling=front views2023.03 | 21.63 | 15.55 | 19.19 | |
| PIZZAReference Source=1 ref. image, Ref. image sampling=all views2023.03 | 20.05 | 15.9 | 18.39 |