Object Pose Estimation on Occluded-LINEMOD (Split #3 Unseen)
96.3Acc15 Accuracycontext-aware weight aggregation network
Evaluation results
| Method | Links | |
|---|---|---|
| context-aware weight aggregation networkBackbone=SD-V1-5, Aggregation=CWA2024.03 | 96.3 | |
| nonlinear aggregation networkBackbone=SD-V1-5, Aggregation=NA2024.03 | 96 | |
| vanilla aggregation networkBackbone=SD-V1-5, Aggregation=VA2024.03 | 95.6 | |
| context-aware weight aggregation networkBackbone=SD-V2-0, Aggregation=CWA2024.03 | 95.5 | |
| context-aware weight aggregation networkBackbone=OpenCLIP, Aggregation=CWA2024.03 | 91.9 | |
| context-aware weight aggregation networkBackbone=DINOv2, Aggregation=CWA2024.03 | 86.3 | |
| Template-poseBackbone=ResNet2024.03 | 85.3 | |
| Zhao et al.Backbone=ResNet2024.03 | 37.5 | |
| MPLBackbone=MPL2024.03 | 17.7 | |
| Wohlhart et al.Backbone=Wohlhart et al.2024.03 | 7.6 | |
| Balntas et al.Backbone=Wohlhart et al.2024.03 | 5.1 |