Object Affordance Labeling on CAD-120
91.8Micro F1 Scorefull model
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| full modelTemporal segmentation=Ground-truth2012.10 | 91.8 | 90.4 | 74.2 | |
| no object interactionsTemporal segmentation=Ground-truth2012.10 | 88.4 | 75.5 | 63.3 | |
| full model with trackingTemporal segmentation=Ground-truth2012.10 | 88.2 | 74.5 | 64.9 | |
| no temporal interactionsTemporal segmentation=Ground-truth2012.10 | 87 | 79.8 | 66.1 | |
| object onlyTemporal segmentation=Ground-truth2012.10 | 86.9 | 72.7 | 63.1 | |
| full, multi-seg learningTemporal segmentation=None (Estimated)2012.10 | 83.9 | 75.9 | 64.2 | |
| full, 1 segment. (best)Temporal segmentation=None (Estimated)2012.10 | 83.1 | 70.1 | 63.9 | |
| full, 1 segment. (averaged)Temporal segmentation=None (Estimated)2012.10 | 81.3 | 67.8 | 60 | |
| full, multi-seg learning + trackingTemporal segmentation=None (Estimated)2012.10 | 79.4 | 62.5 | 50.2 | |
| SVM multiclassTemporal segmentation=Ground-truth2012.10 | 75.6 | 40.6 | 37.9 | |
| image onlyTemporal segmentation=Ground-truth2012.10 | 74.2 | 15.9 | 16 | |
| max classTemporal segmentation=Ground-truth2012.10 | 65.7 | 65.7 | 8.3 |