Volumetric prediction on SURREAL (full)
65.8Avg SMPL Surface ErrorBodyNet
Evaluation Results
| Method | Links | |
|---|---|---|
| BodyNetArchitecture=End-to-end, Intermediate tasks=with, View=FV+SV2018.04 | 65.8 | |
| BodyNetArchitecture=End-to-end, Intermediate tasks=with, View=FV2018.04 | 67.7 | |
| BodyNetArchitecture=Voxels network, Re-projection loss=FV+SV silhouette2018.04 | 68.2 | |
| BodyNetArchitecture=Voxels network, Re-projection loss=FV silhouette2018.04 | 69.9 | |
| BodyNetArchitecture=End-to-end, Intermediate tasks=without, View=FV+SV2018.04 | 70.5 | |
| BodyNetArchitecture=End-to-end, Intermediate tasks=without, View=FV2018.04 | 72.7 | |
| BodyNetArchitecture=Voxels network2018.04 | 73.6 | |
| Shape parameter regressionRegression strategy=beta regressed, theta fixed2018.04 | 74.3 | |
| Tung et al.Input Modality=GT 2D pose and segmentation2018.04 | 74.5 | |
| SMPLify++Optimization strategy=theta, beta optimized2018.04 | 75.3 |