Research map / 3D vision / Human pose estimation
Estimating 3D Motion and Forces of Human–Object Interactions from Internet Videos
Zongmian Li, Jiří Sedlář, Justin Carpentier, Ivan Laptev, Nicolas Mansard, Josef Šivic
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 the 3D data term (ablation) | generic model · Parkour | MPJPE (mm): 117.55 -> 116.13 | Improved |
| add the 3D data term (ablation) | generic model · Parkour (Muscle-up) | MPJPE (mm): 124.12 -> 124.55 | Hurt |
| add force magnitude regularization on top of temporal smoothing (ablation) | generic model · Parkour (L. sole moment) | error (N.m): 23.78 -> 22.91 | Improved |
| add temporal and magnitude force regularization (ablation) | generic model · Parkour (L. sole moment) | error (N.m): 86.54 -> 22.91 | Improved |
| add temporal and magnitude force regularization (ablation) | generic model · Parkour (L. sole linear force) | error (N): 148.74 -> 142.11 | Improved |
| add force magnitude regularization (ablation) | generic model · Parkour (L. hand linear force) | error (N): 109.10 -> 105.07 | Improved |
| use action-specific hyperparameters instead of generic ones | Ours · Parkour | MPJPE (mm): 116.13 -> 115.45 | Improved |
| use action-specific hyperparameters instead of generic ones | Ours · Handtool | MPJPE (mm): 97.51 -> 94.09 | Improved |
| model person-object contact dynamics vs HMR single-frame baseline | Ours (generic model) · Parkour | MPJPE (mm): 135.65 -> 116.13 | Improved |
| model person-object contact dynamics vs HMR single-frame baseline | Ours (generic model) · Handtool | MPJPE (mm): 103.49 -> 97.51 | Improved |
| model person-object interaction vs Mask R-CNN endpoints | Ours (generic model) · Handtool barbell | % within 25 px: 33 -> 47 | Improved |
| model person-object interaction vs Mask R-CNN endpoints (scythe modelled as line segment) | Ours (generic model) · Handtool scythe | % within 25 px: 63 -> 51 | Hurt |
Closest papers by meaning
- Estimating 3D Motion and Forces of Person-Object Interactions From Monocular Video (CVPR 2019 · similarity 0.96)
- Contact and Human Dynamics from Monocular Video (ECCV 2020 · similarity 0.90)
- AnyLift: Scaling Motion Reconstruction from Internet Videos via 2D Diffusion (CVPR 2026 · similarity 0.89)
- InterAct: Advancing Large-Scale Versatile 3D Human-Object Interaction Generation (CVPR 2025 · similarity 0.88)
- MotionPRO: Exploring the Role of Pressure in Human MoCap and Beyond (CVPR 2025 · similarity 0.87)
- How Do I Do That? Synthesizing 3D Hand Motion and Contacts for Everyday Interactions (CVPR 2025 · similarity 0.86)
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.
