3D Human Mesh Estimation on 3DPW (test)
44.1PA-MPJPERemoCap
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| RemoCapApproach category=Video-based Model-Free, Post-processing=false2024.05 | 44.1 | 72.7 | 81.9 | — | — | — | |
| METROTraining sets=H36M+UP3D+MuCo+3DPW+MPII+[COCO]Ours2023.04 | 45.8 | — | — | 17.9 | 17.2 | 31.3 | |
| RemoCapApproach category=Video-based Model-Free, Post-processing=true2024.05 | 45.9 | 74.5 | — | — | — | — | |
| I2L-MeshNetTraining sets=H36M+MuCo+[COCO]Ours2023.04 | 47.1 | — | — | 17.1 | 18.5 | 32.6 | |
| 3DCrowdNetTraining sets=H36M+MuCo+CrowdPose+[COCO+MPII]Ours2023.04 | 47.2 | — | — | 16.8 | 17.7 | 33.5 | |
| PARETraining sets=[COCO]Ours2023.04 | 47.3 | — | — | 17.5 | 18.2 | 32.9 | |
| METROTraining sets=H36M+UP3D+MuCo+3DPW+MPII+[COCO]SMPLify-X2023.04 | 47.9 | — | — | 18.8 | 18.5 | 32.4 | |
| Pose2PoseTraining sets=H36M+[COCO+MPII]Ours2023.04 | 49.6 | — | — | 18.4 | 18.8 | 34.7 | |
| GLOTApproach category=Video-based2024.05 | 50.6 | 80.7 | 96.3 | — | — | — | |
| PyMAFTraining sets=H36M+MI+[COCO+MPII+LSPET]Ours2023.04 | 50.9 | — | — | 18 | 20.8 | 35.1 | |
| CycleAdaptSource dataset=SURREAL [41], 2D pose estimator=OpenPose [3]2023.08 | 51.1 | 84.4 | 99.9 | — | — | — | |
| 3DCrowdNetTrain set=No 3DPW train set2021.04 | 51.5 | 81.7 | 98.3 | — | — | — | |
| 3DCrowdNetTraining sets=H36M+MuCo+CrowdPose+[COCO+MPII]NeuralAnnot2023.04 | 51.5 | — | — | 17.6 | 18.2 | 36.3 | |
| SPINTraining sets=H36M+MI+[COCO+MPII+LSPET]Ours2023.04 | 51.6 | — | — | 19 | 20.7 | 35.4 | |
| MPS-NetApproach category=Video-based2024.05 | 52.1 | 84.3 | 99.7 | — | — | — | |
| TCMRApproach category=Video-based2024.05 | 52.7 | 86.5 | 102.9 | — | — | — | |
| Pose2PoseTraining sets=H36M+[COCO+MPII]Neural Annot2023.04 | 54.4 | — | — | 19.1 | 20.3 | 39 | |
| MEVAApproach category=Video-based2024.05 | 54.7 | 86.9 | — | — | — | — | |
| Fang et al.Train set=No 3DPW train set2021.04 | 54.8 | 85.1 | — | — | — | — | |
| ROMPTrain set=No 3DPW train set2021.04 | 54.9 | 91.3 | 108.3 | — | — | — | |
| CycleAdaptSource Dataset=SURREAL, Adaptation Scenario=Online, 2D Keypoints Source=OpenPose2023.08 | 55.1 | 90 | 106.8 | — | — | — | |
| CycleAdaptSource Dataset=Human3.6M, Adaptation Scenario=Online, 2D Keypoint Detector=OpenPose2023.08 | 55.2 | 90.3 | 107 | — | — | — | |
| TUCHTrain set=No 3DPW train set2021.04 | 55.5 | 84.9 | — | — | — | — | |
| Song et al.Train set=No 3DPW train set2021.04 | 55.9 | — | — | — | — | — | |
| Pose2MeshTrain set=No 3DPW train set2021.04 | 56.3 | 89.5 | 105.3 | — | — | — | |
| PARETraining sets=[COCO]EFT2023.04 | 57.3 | — | — | 20.3 | 20.3 | 41.7 | |
| VIBEApproach category=Video-based2024.05 | 57.6 | 91.9 | — | — | — | — | |
| I2L-MeshNetTrain set=No 3DPW train set2021.04 | 57.7 | 93.2 | 110.1 | — | — | — | |
| I2L-MeshNetTraining sets=H36M+MuCo+[COCO]SMPLify-X2023.04 | 57.7 | — | — | 20.6 | 21.7 | 40.8 | |
| PyMAFTraining sets=H36M+MI+[COCO+MPII+LSPET]SPIN2023.04 | 58.9 | — | — | 21 | 23.7 | 41.8 | |
| SPINTrain set=No 3DPW train set2021.04 | 59.2 | 96.9 | 116.4 | — | — | — | |
| SPINTraining sets=H36M+MI+[COCO+MPII+LSPET]SMPLify2023.04 | 59.6 | — | — | 21.6 | 24.2 | 40.8 | |
| BOASource dataset=SURREAL [41], 2D pose estimator=OpenPose [3]2023.08 | 61.7 | 102.5 | 124.7 | — | — | — | |
| DAPASource dataset=SURREAL [41], 2D pose estimator=OpenPose [3]2023.08 | 61.7 | 96.6 | 122.8 | — | — | — | |
| BOASource Dataset=SURREAL, Adaptation Scenario=Online, 2D Keypoints Source=OpenPose2023.08 | 61.7 | 102.5 | 124.7 | — | — | — | |
| DynaBOASource dataset=SURREAL [41], 2D pose estimator=OpenPose [3]2023.08 | 62.4 | 109.8 | 139.9 | — | — | — | |
| DynaBOASource Dataset=SURREAL, Adaptation Scenario=Online, 2D Keypoints Source=OpenPose2023.08 | 62.4 | 109.8 | 139.9 | — | — | — | |
| GraphCMRTrain set=No 3DPW train set2021.04 | 70.2 | — | — | — | — | — | |
| DynaBOASource Dataset=Human3.6M, Adaptation Scenario=Online, 2D Keypoint Detector=OpenPose2023.08 | 73 | 135.1 | 168.2 | — | — | — | |
| BOASource Dataset=Human3.6M, Adaptation Scenario=Online, 2D Keypoint Detector=OpenPose2023.08 | 76.2 | 137.6 | 171.8 | — | — | — | |
| HMRTrain set=No 3DPW train set2021.04 | 76.7 | 130 | — | — | — | — | |
| Base modelPre-trained on=SURREAL [41]2023.08 | 92 | 193.2 | 216.5 | — | — | — | |
| Base modelSource Dataset=SURREAL, Adaptation Scenario=Online, 2D Keypoints Source=OpenPose2023.08 | 92 | 193.2 | 216.5 | — | — | — | |
| Base model (pre-trained on H36M)Source Dataset=Human3.6M, Adaptation Scenario=Online2023.08 | 123.4 | 230.3 | 253.4 | — | — | — |