3D Human Pose and Shape Estimation on 3DPW
32.8PA-MPJPECLIFF
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| CLIFFRole=pseudo-GT annotator2022.08 | 32.8 | 52.8 | 61.5 | — | — | — | — | — | |
| Pose2MeshRole=pseudo-GT annotator2022.08 | 34.6 | 65.1 | — | — | — | — | — | — | |
| Ours2022.10 | 39 | 66 | 76.3 | — | — | — | — | — | |
| DynaBOARole=pseudo-GT annotator2022.08 | 40.4 | 65.5 | 82 | — | — | — | — | — | |
| CLIFFMethod Category=model-based, Backbone=HRNet-W482022.08 | 43 | 69 | 81.2 | — | — | — | — | — | |
| IKOLBackbone=ResNet-34, Fine-tuning=fine tuned with the HybrIK model2023.02 | 44.5 | 71.1 | 84.1 | — | — | — | — | — | |
| IKOLBackbone=ResNet-34, Fine-tuning=false2023.02 | 45.5 | 73.3 | 86.4 | — | — | — | — | — | |
| GraphormerMethod Category=model-free2022.08 | 45.6 | 74.7 | 87.7 | — | — | — | — | — | |
| Mesh Graphormer2022.10 | 45.6 | 74.7 | 87.7 | — | — | — | — | — | |
| Mesh Graphormer2023.02 | 45.6 | 74.7 | 87.7 | — | — | — | — | — | |
| MAEDMethod Category=video2022.08 | 45.7 | 79.1 | 92.6 | — | — | — | — | — | |
| CLIFFMethod Category=model-based, Backbone=ResNet-502022.08 | 45.7 | 72 | 85.3 | — | — | — | — | — | |
| PAREMethod Category=model-based2022.08 | 46.4 | 79.1 | 94.2 | — | — | — | — | — | |
| PAREBackbone=HRNet-w322023.02 | 46.4 | 79.1 | 94.2 | — | — | — | — | — | |
| MeshLeTemp2022.10 | 46.8 | 74.8 | 86.5 | — | — | — | — | — | |
| BEVBackbone=HRNet-w322023.02 | 46.9 | 78.5 | 92.3 | — | — | — | — | — | |
| ROMPMethod Category=model-based2022.08 | 47.3 | 76.7 | 93.4 | — | — | — | — | — | |
| ROMPBackbone=HRNet-w322023.02 | 47.3 | 76.7 | 93.4 | — | — | — | — | — | |
| METROMethod Category=model-free2022.08 | 47.9 | 77.1 | 88.2 | — | — | — | — | — | |
| METRO2022.10 | 47.9 | 77.1 | 88.2 | — | — | — | — | — | |
| METROBackbone=HRNet-w642023.02 | 47.9 | 77.1 | 88.2 | — | — | — | — | — | |
| HybrIKMethod Category=model-free2022.08 | 48.8 | 80 | 94.5 | — | — | — | — | — | |
| HybrIK2022.10 | 48.8 | 80 | 94.5 | — | — | — | — | — | |
| HybrIKBackbone=ResNet-342023.02 | 48.8 | 80 | 94.5 | — | — | — | — | — | |
| EFTRole=pseudo-GT annotator2022.08 | 49.3 | — | — | — | — | — | — | — | |
| BOARole=pseudo-GT annotator2022.08 | 49.5 | 77.2 | — | — | — | — | — | — | |
| GLoTnumber of input frames=162023.03 | 50.6 | 80.7 | — | 96.3 | 6.6 | — | — | — | |
| PAREBackbone=HRNet-322022.10 | 50.9 | 82 | 97.9 | — | — | — | — | — | |
| VIBEMethod Category=video2022.08 | 51.9 | 82.7 | 99.1 | — | — | — | — | — | |
| MPS-Netnumber of input frames=162023.03 | 52.1 | 84.3 | — | 99.7 | 7.4 | — | — | — | |
| HMR-EFTMethod Category=model-based2022.08 | 52.2 | 85.1 | 98.7 | — | — | — | — | — | |
| 3DCrowdNet2022.10 | 52.2 | 82.8 | 100.2 | — | — | — | — | — | |
| HMR-EFTBackbone=ResNet-502023.02 | 52.2 | 85.1 | 98.7 | — | — | — | — | — | |
| PAREBackbone=Resnet-502022.10 | 52.3 | 82.9 | 99.7 | — | — | — | — | — | |
| ProHMRRole=pseudo-GT annotator2022.08 | 52.4 | — | — | — | — | — | — | — | |
| TCMRMethod Category=video2022.08 | 52.7 | 86.5 | 102.9 | — | — | — | — | — | |
| TCMRnumber of input frames=162023.03 | 52.7 | 86.5 | — | 102.9 | 7.1 | — | — | — | |
| SPECMethod Category=model-based2022.08 | 53.2 | 96.5 | 118.5 | — | — | — | — | — | |
| SPECBackbone=ResNet-502023.02 | 53.2 | 96.5 | 118.5 | — | — | — | — | — | |
| ROMPBackbone=HRNet-322022.10 | 53.3 | 85.5 | 103.1 | — | — | — | — | — | |
| ROMPBackbone=Resnet-502022.10 | 53.5 | 89.3 | 105.6 | — | — | — | — | — | |
| MEVAnumber of input frames=902023.03 | 54.7 | 86.9 | — | — | 11.6 | — | — | — | |
| Song et al.Category=Frame-based2021.03 | 55.9 | — | — | — | — | — | — | — | |
| Pose2MeshMethod Category=model-free2022.08 | 56.3 | 89.5 | 105.3 | — | — | — | — | — | |
| VIBE2022.10 | 56.5 | 93.5 | 116.4 | — | — | — | — | — | |
| VIBECategory=Temporal2021.03 | 56.5 | 93.5 | 113.4 | — | — | — | — | — | |
| DaNetCategory=Frame-based2021.03 | 56.9 | — | — | — | — | — | — | — | |
| VIBEnumber of input frames=162023.03 | 57.6 | 91.9 | — | — | 25.4 | — | — | — | |
| I2L-MeshNetMethod Category=model-free2022.08 | 58.6 | 93.2 | 110.1 | — | — | — | — | — | |
| I2L-MeshNetBackbone=ResNet-502023.02 | 58.6 | 93.2 | 110.1 | — | — | — | — | — | |
| PyMAFCategory=Frame-based, Auxiliary supervision (AS)=false2021.03 | 58.8 | 95.6 | 113.6 | — | — | — | — | — | |
| PyMAFCategory=Frame-based, Auxiliary supervision (AS)=true2021.03 | 58.9 | 92.8 | 110.1 | — | — | — | — | — | |
| SPINMethod Category=model-based2022.08 | 59.2 | 96.9 | 116.4 | — | — | — | — | — | |
| SPIN2022.10 | 59.2 | 96.6 | 116.4 | — | — | — | — | — | |
| SPINBackbone=ResNet-502023.02 | 59.2 | 96.9 | 116.4 | — | — | — | — | — | |
| SPINCategory=Frame-based2021.03 | 59.2 | 96.9 | 116.4 | — | — | — | — | — | |
| HKMRCategory=Frame-based2021.03 | 59.6 | — | — | — | — | — | — | — | |
| I2L-MeshNetCategory=Frame-based2021.03 | 60 | 100 | — | — | — | — | — | — | |
| BaselineCategory=Frame-based2021.03 | 60.9 | 98.5 | 117.9 | — | — | — | — | — | |
| DecoMRCategory=Frame-based, Non-parametric evaluation=true2021.03 | 61.7 | — | — | — | — | — | — | — | |
| BMP2022.10 | 63.8 | 104.1 | 119.3 | — | — | — | — | — | |
| DSDCategory=Temporal2021.03 | 69.5 | — | — | — | — | — | — | — | |
| Kolotouros et al.Backbone=ResNet-502023.02 | 70.2 | — | — | — | — | — | — | — | |
| GraphCMRCategory=Frame-based2021.03 | 70.2 | — | — | — | — | — | — | — | |
| Arnab et al.2023.02 | 72.2 | — | — | — | — | — | — | — | |
| Arnab et al.Category=Temporal2021.03 | 72.2 | — | — | — | — | — | — | — | |
| HMMRMethod Category=video2022.08 | 72.6 | 116.5 | — | — | — | — | — | — | |
| temporal HMR2022.10 | 72.6 | 116.5 | 139.3 | — | — | — | — | — | |
| Kanazawa et al.Category=Temporal2021.03 | 72.6 | 116.5 | 139.3 | — | — | — | — | — | |
| Doersch et al.Category=Temporal2021.03 | 74.7 | — | — | — | — | — | — | — | |
| HMR2022.10 | 76.7 | 130 | — | — | — | — | — | — | |
| HMRCategory=Frame-based2021.03 | 76.7 | 130 | — | — | — | — | — | — | |
| HMRMethod Category=model-based2022.08 | 81.3 | 130 | — | — | — | — | — | — | |
| HMRBackbone=ResNet-502023.02 | 81.3 | 130 | — | — | — | — | — | — | |
| ACAModel=SMPLer-X-H2025.05 | — | — | — | — | — | 77.3 | 52.47 | 3.05 | |
| ACAModel=SMPLer-X-L2025.05 | — | — | — | — | — | 78.53 | 53.02 | 3.53 | |
| ACAModel=SMPLer-X-B2025.05 | — | — | — | — | — | 82.68 | 55.6 | 4.05 | |
| ACAModel=SMPLer-X-S2025.05 | — | — | — | — | — | 85.52 | 58.86 | 3.45 | |
| ACAModel=OSX2025.05 | — | — | — | — | — | 98.05 | 66.39 | 3.95 | |
| ACAModel=Hand4Whole2025.05 | — | — | — | — | — | 100.7 | 67.95 | 0.04 | |
| CleanModel=SMPLer-X-H2025.05 | — | — | — | — | — | 75.01 | 50.57 | — | |
| CleanModel=SMPLer-X-L2025.05 | — | — | — | — | — | 75.85 | 50.67 | — | |
| CleanModel=SMPLer-X-B2025.05 | — | — | — | — | — | 79.46 | 52.62 | — | |
| CleanModel=SMPLer-X-S2025.05 | — | — | — | — | — | 82.67 | 56.65 | — | |
| CleanModel=OSX2025.05 | — | — | — | — | — | 94.32 | 63.87 | — | |
| CleanModel=Hand4Whole2025.05 | — | — | — | — | — | 100.66 | 67.93 | — | |
| DiffAttackModel=SMPLer-X-H2025.05 | — | — | — | — | — | 80.46 | 54.26 | 7.27 | |
| DiffAttackModel=SMPLer-X-L2025.05 | — | — | — | — | — | 78.7 | 53.25 | 3.76 | |
| DiffAttackModel=SMPLer-X-B2025.05 | — | — | — | — | — | 82.93 | 56.31 | 4.37 | |
| DiffAttackModel=SMPLer-X-S2025.05 | — | — | — | — | — | 85.61 | 58.97 | 3.56 | |
| DiffAttackModel=OSX2025.05 | — | — | — | — | — | 97.78 | 66.1 | 3.67 | |
| DiffAttackModel=Hand4Whole2025.05 | — | — | — | — | — | 100.54 | 67.68 | -0.12 | |
| FGSMModel=SMPLer-X-H2025.05 | — | — | — | — | — | 84.4 | 56.04 | 12.52 | |
| FGSMModel=SMPLer-X-L2025.05 | — | — | — | — | — | 85.05 | 56.2 | 12.13 | |
| FGSMModel=SMPLer-X-B2025.05 | — | — | — | — | — | 94.04 | 60.97 | 18.35 | |
| FGSMModel=SMPLer-X-S2025.05 | — | — | — | — | — | 100.05 | 65.32 | 21.02 | |
| FGSMModel=OSX2025.05 | — | — | — | — | — | 102.09 | 68.13 | 8.24 | |
| FGSMModel=Hand4Whole2025.05 | — | — | — | — | — | 100.66 | 67.93 | 0 | |
| LatentStealthModel=SMPLer-X-H2025.05 | — | — | — | — | — | 103.45 | 65.4 | 37.91 | |
| LatentStealthModel=SMPLer-X-L2025.05 | — | — | — | — | — | 104.53 | 64.94 | 37.81 |