Human Mesh Recovery on Human3.6M
40.2Reconstruction ErrorKAMA
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| KAMAMesh Supervision=N, Optimization-based=true, description=Ours w. refinement2021.04 | 40.2 | — | — | |
| SPIN et al.Mesh Supervision=Y, Optimization-based=false2021.04 | 41.1 | — | — | |
| ProHMRInput=image, T=12022.10 | 41.2 | — | — | |
| KAMAMesh Supervision=N, Optimization-based=false2021.04 | 41.5 | — | — | |
| I2L-MeshNetMesh Supervision=Y, Optimization-based=false2021.04 | 41.7 | — | — | |
| DSDMesh Supervision=Y, Optimization-based=false2021.04 | 44.3 | — | — | |
| HoloPoseMesh Supervision=Y, Optimization-based=false2021.04 | 46.5 | — | — | |
| Pose2MeshMesh Supervision=Y, Optimization-based=false2021.04 | 47 | — | — | |
| CLIFFInput=image, T=12022.10 | 47.1 | — | — | |
| Kundu et al.Mesh Supervision=N, Optimization-based=false2021.04 | 48.1 | — | — | |
| TexturePoseMesh Supervision=Y, Optimization-based=false2021.04 | 49.7 | — | — | |
| GraphCMRMesh Supervision=Y, Optimization-based=false2021.04 | 50.1 | — | — | |
| VisDBInput=image, T=12022.10 | 51 | — | — | |
| DMAPS2026.02 | 51.18 | 35.96 | 3.92 | |
| Mesh GraphormerInput=image, T=12022.10 | 51.2 | — | — | |
| ARTS2026.02 | 51.55 | 36.63 | 4.11 | |
| FastMETROInput=image, T=12022.10 | 52.2 | — | — | |
| MotionBERT (finetune) + HybrIKInput=video, T=162022.10 | 52.6 | — | — | |
| METROInput=image, T=12022.10 | 54 | — | — | |
| PMCE2026.02 | 54.58 | 37.26 | 3.02 | |
| STRAPSMesh Supervision=N, Optimization-based=false2021.04 | 55.4 | — | — | |
| I2L-MeshNetInput=image, T=12022.10 | 55.7 | — | — | |
| Song et al.Mesh Supervision=N, Optimization-based=true2021.04 | 56.4 | — | — | |
| HMRMesh Supervision=Y, Optimization-based=false2021.04 | 56.8 | — | — | |
| HMMRInput=video, T=202022.10 | 56.9 | — | — | |
| PoseBERTInput=video, T=162022.10 | 57.3 | — | — | |
| MEAD2026.02 | 57.64 | 39.08 | 5.86 | |
| PyMAFInput=image, T=12022.10 | 57.7 | — | — | |
| HybrIKInput=image, T=12022.10 | 58.1 | — | — | |
| DSD-SATNInput=video, T=92022.10 | 59.1 | — | — | |
| TCMRInput=video, T=162022.10 | 62.3 | — | — | |
| MotionBERT (finetune) + SPINInput=video, T=162022.10 | 63.7 | — | — | |
| MotionBERT (finetune)Input=2D motion, T=162022.10 | 65.5 | — | — | |
| VIBEInput=video, T=162022.10 | 65.6 | — | — | |
| GLoT2026.02 | 66.05 | 45.74 | 4.68 | |
| MotionBERT (finetune) + MAEDInput=video, T=162022.10 | 66.8 | — | — | |
| SmoothNetInput=video, T=322022.10 | 67.5 | — | — | |
| MPS-NetInput=video, T=162022.10 | 69.4 | — | — | |
| MotionBERT (scratch)Input=2D motion, T=162022.10 | 75.7 | — | — | |
| SMPLify-XMesh Supervision=N, Optimization-based=true2021.04 | 75.9 | — | — | |
| VIBE2026.02 | 77.24 | 54.06 | 26.72 | |
| TemporalContextInput=video, T=322022.10 | 77.8 | — | — | |
| SPINInput=image, T=12022.10 | 82.3 | — | — | |
| SMPLifyMesh Supervision=N, Optimization-based=true2021.04 | 82.3 | — | — | |
| MAEDInput=video, T=162022.10 | 84.1 | — | — | |
| Pose2MeshInput=2D pose, T=12022.10 | 85.3 | — | — | |
| HMRInput=image, T=12022.10 | 88 | — | — | |
| Biggs et al.2021.08 | — | 41.6 | — | |
| Biggs et al. (CVAE)Number of hypotheses (n)=52021.08 | — | 46.4 | — | |
| Biggs et al. (CVAE)Number of hypotheses (n)=102021.08 | — | 46.3 | — | |
| Biggs et al. (CVAE)Number of hypotheses (n)=252021.08 | — | 46.2 | — | |
| Biggs et al. (CVAE)Number of hypotheses (n)=min2021.08 | — | 46.2 | — | |
| Biggs et al. (MDN)Number of hypotheses (n)=52021.08 | — | 43.3 | — | |
| Biggs et al. (MDN)Number of hypotheses (n)=102021.08 | — | 43 | — | |
| Biggs et al. (MDN)Number of hypotheses (n)=252021.08 | — | 42.7 | — | |
| Biggs et al. (MDN)Number of hypotheses (n)=min2021.08 | — | 42.7 | — | |
| Biggs et al. (NF)Number of hypotheses (n)=52021.08 | — | 42 | — | |
| Biggs et al. (NF)Number of hypotheses (n)=102021.08 | — | 42.2 | — | |
| Biggs et al. (NF)Number of hypotheses (n)=252021.08 | — | 42.2 | — | |
| Biggs et al. (NF)Number of hypotheses (n)=min2021.08 | — | 41.6 | — | |
| HMR2021.08 | — | 56.8 | — | |
| ProHMRNumber of hypotheses (n)=52021.08 | — | 39.4 | — | |
| ProHMRNumber of hypotheses (n)=102021.08 | — | 38.3 | — | |
| ProHMRNumber of hypotheses (n)=252021.08 | — | 36.8 | — | |
| ProHMRNumber of hypotheses (n)=min2021.08 | — | 29.9 | — | |
| ProHMR2021.08 | — | 41.2 | — | |
| SPIN2021.08 | — | 41.1 | — |