Research map / 3D vision / Human pose estimation
Human Mesh Recovery From Multiple Shots
Georgios Pavlakos, Jitendra Malik, Angjoo Kanazawa
Check 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 MS-AVA training data to HMR | HMR single-frame model · Partial Humans Uncropped (VLOG) | PCKh: 86.9 -> 90.3 | Improved |
| Add MS-AVA training data to HMR | HMR single-frame model · Partial Humans Uncropped (YouCook) | PCKh: 96.8 -> 98.9 | Improved |
| Add MS-AVA training data to HMR | HMR single-frame model · Partial Humans Uncropped (Instr) | PCKh: 92.4 -> 94.1 | Improved |
| Add MS-AVA training data to HMR | HMR single-frame model · Partial Humans Uncropped (Cross) | PCKh: 96.2 -> 98.2 | Improved |
| Add MS-AVA training data to HMR | HMR single-frame model · 3DPW | PA-MPJPE (mm): 59.2 -> 57.8 | Improved |
| Replace the recurrent temporal encoder with a transformer | Temporal human mesh recovery · 3DPW test | PA-MPJPE (mm): 56.5 -> 56.1 | Improved |
| Use a transformer instead of convolutional or recurrent temporal encoding | Temporal human mesh recovery · AVA (train: clean, test: clean) | cross-shot PCK: 68.8; 70.8; 70.7 | No clear effect |
| Train with sequences containing missing frames and use a transformer temporal encoder | Temporal human mesh recovery · AVA (train: all, test: clean) | cross-shot PCK: 68.8 -> 73.8 | Improved |
| Use a transformer temporal encoder on sequences with missing frames | Temporal human mesh recovery · AVA (train: all, test: all) | cross-shot PCK: 67.5; 68.5; 70.7 | Improved |
| Use a transformer temporal encoder rather than convolutional or recurrent encoding | Temporal human mesh recovery · VIBE training setting | training stability | Improved |
| Use multi-shot optimization rather than single-frame or single-shot optimization | 3D human mesh reconstruction · AVA cross-shot evaluation | cross-shot PCK | 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.
- VIBE — VIBE: Video Inference for Human Body Pose and Shape Estimation (CVPR 2020) · better on 1 of 1 shared result
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
- Full-Body Awareness from Partial Observations (ECCV 2020 · similarity 0.89)
- Reconstructing People, Places, and Cameras (CVPR 2025 · similarity 0.88)
- FactorizedHMR: A Hybrid Framework for Video Human Mesh Recovery (NeurIPS 2026 · similarity 0.88)
- Guess The Unseen: Dynamic 3D Scene Reconstruction from Partial 2D Glimpses (CVPR 2024 · similarity 0.87)
- ODHSR: Online Dense 3D Reconstruction of Humans and Scenes from Monocular Videos (CVPR 2025 · similarity 0.87)
- HumanMM: Global Human Motion Recovery from Multi-shot Videos (CVPR 2025 · similarity 0.87)
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.
