Research map / 3D vision / Human pose estimation
Probabilistic Modeling for Human Mesh Recovery
Nikos Kolotouros, Georgios Pavlakos, Dinesh Jayaraman, Kostas Daniilidis
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add explicit supervision of the distribution mode (L_mode) (ablation) | ProHMR · 3DPW | PA-MPJPE: 67.4 -> 59.8 | Improved |
| add explicit supervision of the distribution mode (L_mode) (ablation) | ProHMR · Human3.6M | PA-MPJPE: 54.8 -> 41.2 | Improved |
| add explicit supervision of the distribution mode (L_mode) (ablation) | ProHMR · MPI-INF-3DHP | PA-MPJPE: 76.5 -> 65.0 | Improved |
| add image-conditioned prior fitting on top of ProHMR regression | ProHMR · 3DPW | PA-MPJPE: 59.8 -> 55.1 | Improved |
| add image-conditioned prior fitting on top of ProHMR regression | ProHMR · Human3.6M GT keypoints | PA-MPJPE: 41.2 -> 34.8 | Improved |
| add SMPLify fitting (VPoser) on top of SPIN | SPIN · 3DPW | PA-MPJPE: 59.2 -> 70.9 | Hurt |
| add EFT fitting on top of SPIN | SPIN · 3DPW | PA-MPJPE: 59.2 -> 56.6 | Improved |
| add multi-view fitting with ProHMR prior instead of single-view prediction | ProHMR · Human3.6M multi-view | PA-MPJPE: 43.7 -> 34.5 | Improved |
| use fitting-based multi-view refinement instead of rotation averaging | ProHMR · Mannequin Challenge | PA-MPJPE: 85.1 -> 83.9 | Improved |
| use normalizing flow instead of Mixture Density Network for the distribution | ProHMR · 3DPW multiple hypotheses (n=5) | PA-MPJPE: 61.2 -> 56.5 | Improved |
| use ProHMR instead of Conditional VAE for the distribution | ProHMR · Human3.6M multiple hypotheses (n=25) | PA-MPJPE: 46.2 -> 36.8 | Improved |
| use ProHMR mode instead of SPIN regression | ProHMR · MPI-INF-3DHP | PA-MPJPE: 67.5 -> 65.0 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SPIN — SPIN: Structure-Preserving Inner Offset Network for Scene Text Recognition (AAAI 2021) · better on 4 of 6 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Self-Supervised 3D Human Mesh Recovery from a Single Image with Uncertainty-Aware Learning (AAAI 2024 · similarity 0.89)
- DiffHuman: Probabilistic Photorealistic 3D Reconstruction of Humans (CVPR 2024 · similarity 0.89)
- GenHMR: Generative Human Mesh Recovery (AAAI 2025 · similarity 0.88)
- TokenHMR: Advancing Human Mesh Recovery with a Tokenized Pose Representation (CVPR 2024 · similarity 0.88)
- Reconstructing Humans with a Biomechanically Accurate Skeleton (CVPR 2025 · similarity 0.88)
- DPMesh: Exploiting Diffusion Prior for Occluded Human Mesh Recovery (CVPR 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
