Research map / 3D vision / 3D generation
Articulat3D: Reconstructing Articulated Digital Twins From Monocular Videos with Geometric and Motion Constraints
Lijun Guo, Haoyu Zhao, Xingyue Zhao, Rong Fu, Linghao Zhuang, Siteng Huang, Zhongyu Li, Hua Zou
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 3D track motion prior (motion bases) guidance (ablation) | Articulat3D · Articulat3D-Sim ablation | axis error (deg): 43.42 -> 0.53 | Improved |
| add motion prior-driven initialization before kinematic refinement (ablation) | Articulat3D · Articulat3D-Sim ablation | PSNR: 30.89 -> 37.80 | Improved |
| add geometric and motion constraints refinement (kinematic primitives) (ablation) | Articulat3D · Articulat3D-Sim ablation | position error (cm): 4.98 -> 0.65 | Improved |
| add kinematic constraints refinement (ablation) | Articulat3D · Articulat3D-Sim ablation | LPIPS: 0.02 -> 0.03 | Mixed |
| train with only 1/3 of frames (uniform subsampling) (ablation) | Articulat3D · Articulat3D-Sim ablation | PSNR: 37.80 -> 34.60 | Hurt |
| train with only 1/3 of frames (ablation) | Articulat3D · Articulat3D-Sim ablation | axis error (deg): 0.53 -> 0.61 | No clear effect |
| inject Gaussian noise into TAPIP3D tracks (ablation) | Articulat3D · Articulat3D-Sim ablation | axis error (deg): 0.53 -> 0.56 | No clear effect |
| use Articulat3D instead of Shape of Motion | dynamic scene reconstruction · Articulat3D-Sim | PSNR: 33.37 -> 37.80 | Improved |
| use Articulat3D instead of VideoArtGS | articulated reconstruction · VideoArtGS-20 | axis error (deg): 0.34 -> 0.27 | Improved |
| use learning rate 1e-3 instead of 1e-4 for joint axis and pivot (ablation) | Articulat3D · Articulat3D-Sim hyperparameters | position error (cm): 0.65 -> 5.82 | Hurt |
Closest papers by meaning
- Articulated Object Reconstruction from Rest-State Observation (ECCV 2026 · similarity 0.92)
- ArticulatedGS: Self-supervised Digital Twin Modeling of Articulated Objects using 3D Gaussian Splatting (CVPR 2025 · similarity 0.90)
- Part$^{2}$GS: Part-aware Modeling of Articulated Objects using 3D Gaussian Splatting (CVPR 2026 · similarity 0.89)
- Articulation in Prime: Primitive-Based Articulated Object Understanding from a Single Casual Video (NeurIPS 2026 · similarity 0.88)
- Building Interactable Replicas of Complex Articulated Objects via Gaussian Splatting (ICLR 2025 · similarity 0.86)
- DIPO: Dual-State Images Controlled Articulated Object Generation Powered by Diverse Data (NeurIPS 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.
