Research map / 3D vision / 3D generation
ScoreHOI: Physically Plausible Reconstruction of Human-Object Interaction via Score-Guided Diffusion
Ao Li, Jinpeng Liu, Yixuan Zhu, Yansong Tang
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 diffusion sampling loop (ablation) | ScoreHOI · BEHAVE | Contact F-Score: 0.588 -> 0.609 | Improved |
| add contact-driven iterative refinement (ablation) | ScoreHOI · BEHAVE | Contact F-Score: 0.577 -> 0.609 | Improved |
| add contact-driven iterative refinement (ablation) | ScoreHOI · BEHAVE | CD_human: 4.93 -> 4.85 | Improved |
| add image and geometry conditions (ablation) | ScoreHOI · BEHAVE | Contact F-Score: 0.585 -> 0.609 | Improved |
| add geometry condition c_G (ablation) | ScoreHOI · BEHAVE | Contact F-Score: 0.591 -> 0.609 | Improved |
| add image condition c_I (ablation) | ScoreHOI · BEHAVE | Contact F-Score: 0.597 -> 0.609 | Improved |
| add physical guidance losses (ablation) | ScoreHOI · BEHAVE | Contact F-Score: 0.570 -> 0.609 | Improved |
| add human-object contact guidance L_ho (ablation) | ScoreHOI · BEHAVE | Contact recall: 0.525 -> 0.586 | Improved |
| add penetration guidance L_pt (ablation) | ScoreHOI · BEHAVE | Contact precision: 0.619 -> 0.634 | Improved |
| add L_of guidance (ablation) | ScoreHOI · BEHAVE | Contact F-Score: 0.602 -> 0.609 | Improved |
| increase DDIM sampling steps (step size 25 to 2) (ablation) | ScoreHOI · BEHAVE | CD_human: 5.80 -> 4.85 | Improved |
| increase iterations N from 2 to 10 (ablation) | ScoreHOI · BEHAVE | Contact F-Score: 0.604 -> 0.609 | Improved |
Closest papers by meaning
- Real2Sim in HOI: Toward Physically Plausible HOI Reconstruction from Monocular Videos (NeurIPS 2026 · similarity 0.87)
- Score-Guided Diffusion for 3D Human Recovery (CVPR 2024 · similarity 0.87)
- Template Free Reconstruction of Human-object Interaction with Procedural Interaction Generation (CVPR 2024 · similarity 0.87)
- CG-HOI: Contact-Guided 3D Human-Object Interaction Generation (CVPR 2024 · similarity 0.87)
- Reconstructing People, Places, and Cameras (CVPR 2025 · similarity 0.86)
- An Image-like Diffusion Method for Human-Object Interaction Detection (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.
