Research map / 3D vision / 3D generation
PhysRig: Differentiable Physics-Based Skinning and Rigging Framework for Realistic Articulated Object Modeling
Hao Zhang, Haolan Xu, Chun Feng, Varun Jampani, Narendra Ahuja
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 |
|---|---|---|---|
| replace LBS with differentiable physics-based skinning (PhysRig) | PhysRig · Inverse skinning, 17-object dataset average | Chamfer Distance: 0.73 -> 0.37 (vs LBS-3) | Improved |
| replace LBS with PhysRig | PhysRig · Inverse skinning, 17-object dataset average | User Study Rate: 3.35 -> 4.52 (vs LBS-3) | Improved |
| run iterative optimization of material parameters and driving point velocities after initialization (ablation) | PhysRig · Inverse skinning, 17-object dataset average | Chamfer Distance: 4.67 -> 0.37 | Improved |
| use material prototypes instead of per-point material parameters (ablation) | PhysRig · Michelle | Chamfer Distance: 2.31 -> 0.139 | Improved |
| use material prototypes instead of triplane material field (ablation) | PhysRig · Michelle | Chamfer Distance: 1.93 -> 0.139 | Improved |
| add joint localization for driving point initialization (ablation) | PhysRig · Michelle | convergence iterations: 8000 -> 2500 | Improved |
| add velocity initialization for driving points (ablation) | PhysRig · Michelle | convergence iterations: 5000 -> 2500 | Improved |
| reduce material prototypes from 100 to 25 (ablation) | PhysRig · Michelle | convergence iterations: 2500 -> 2000 | Improved |
| increase material prototypes from 100 to 200 (ablation) | PhysRig · Michelle | Chamfer Distance: 0.139 -> 0.133 | No clear effect |
| replace LBS with PhysRig on a case with heavy deformation (Jellyman) | PhysRig · Jellyman | Chamfer Distance: 3.811 -> 1.214 (vs LBS-3) | Improved |
Closest papers by meaning
- REACTO: Reconstructing Articulated Objects from a Single Video (CVPR 2024 · similarity 0.88)
- REArtGS: Reconstructing and Generating Articulated Objects via 3D Gaussian Splatting with Geometric and Motion Constraints (NeurIPS 2025 · similarity 0.86)
- RigGS: Rigging of 3D Gaussians for Modeling Articulated Objects in Videos (CVPR 2025 · similarity 0.86)
- MagicArticulate: Make Your 3D Models Articulation-Ready (CVPR 2025 · similarity 0.86)
- MPMAvatar: Learning 3D Gaussian Avatars with Accurate and Robust Physics-Based Dynamics (NeurIPS 2025 · similarity 0.86)
- Learning Implicit Representation for Reconstructing Articulated Objects (ICLR 2024 · similarity 0.85)
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.
