3D Human Pose Estimation on HumanEva-I (S1 Box)
117.1P[G] Global3DPoseNet
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| 3DPoseNetInput Modality=Monocular, GT Bounding Box Usage=false, Trained/Fine-tuned on HumanEva-I=false2016.11 | 117.1 | 80.5 | 58.6 | |
| 3DPoseNet (w/o Persp. correct.)Input Modality=Monocular, GT Bounding Box Usage=false, Trained/Fine-tuned on HumanEva-I=false, Perspective Correction=false2016.11 | 116.1 | 79.4 | 58.6 | |
| 3DPoseNetInput Modality=Monocular, GT Bounding Box Usage=false, Trained/Fine-tuned on HumanEva-I=true2016.11 | 77.9 | 38.4 | 30.8 | |
| Elhayek et al. [18]Input Modality=Multi-view2016.11 | 60 | — | — | |
| Rhodin et al. [50]Input Modality=Multi-view2016.11 | 59.7 | — | — | |
| Amin et al. [3]Input Modality=Multi-view2016.11 | 47.7 | — | — | |
| Akhter et al. [2]Input Modality=Monocular2016.11 | — | — | 165.5 | |
| Bogo et al. [9]Input Modality=Monocular, Trained/Fine-tuned on HumanEva-I=true2016.11 | — | — | 82.1 | |
| Ramakris. et al. [48]Input Modality=Monocular2016.11 | — | — | 151 |