Research map / 3D vision / 3D generation
DAViD: Modeling Dynamic Affordance of 3D Objects Using Pre-trained Video Diffusion Models
Hyeonwoo Kim, Sangwon Baik, Hanbyul Joo
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 contact guidance in TGS (ablation) | DAViD · FullBodyManipulation (CHOIS split) | Contact recall: 0.154 -> 0.577 | Improved |
| add contact guidance in TGS (ablation) | DAViD · FullBodyManipulation (CHOIS split) | Contact F1: 0.264 -> 0.698 | Improved |
| add contact guidance in TGS (ablation) | DAViD · FullBodyManipulation (OMOMO split) | Contact F1: 0.302 -> 0.667 | Improved |
| add contact guidance in TGS (ablation) | DAViD · FullBodyManipulation (CHOIS split) | Contact precision: 0.923 -> 0.882 | Hurt |
| add smoothness guidance in TGS (ablation) | DAViD · FullBodyManipulation | Object jitter: 2.95 x 10^3 -> 30.9 | Improved |
| add LoRA to MDM (ablation) | DAViD · HumanML3D | FID: 2.93 -> 1.30 | Improved |
| add LoRA to MDM (ablation) | DAViD · HumanML3D | Diversity: 8.59 -> 8.90 | Improved |
| use DAViD instead of CHOIS | DAViD · FullBodyManipulation | Contact F1: 0.565 -> 0.698 | Improved |
| use DAViD instead of OMOMO | DAViD · FullBodyManipulation | Contact recall: 0.547 -> 0.558 | No clear effect |
| use DAViD instead of CHOIS | DAViD · FullBodyManipulation | Foot sliding: 0.472 -> 0.234 | Improved |
| raise potential-contact threshold from 0.00 to 0.02 (ablation) | DAViD · FullBodyManipulation (CHOIS split) | Contact F1 (average): 0.524 -> 0.634 | Improved |
| raise potential-contact threshold from 0.02 to 0.03 (ablation) | DAViD · FullBodyManipulation (CHOIS split) | Contact F1 (average): 0.634 -> 0.614 | Hurt |
Later papers that compared with this method
2 later accepted papers report DAViD as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- Skullptor: High Fidelity 3D Head Reconstruction in Seconds with Multi-View Normal Prediction (CVPR 2026 · better on 4 of 4 shared results)
- Face Anything: 4D Face Reconstruction from Any Image Sequence (ECCV 2026 · 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
- HOI-Dyn: Learning Interaction Dynamics for Human-Object Motion Diffusion (NeurIPS 2025 · similarity 0.92)
- InterAct: Advancing Large-Scale Versatile 3D Human-Object Interaction Generation (CVPR 2025 · similarity 0.91)
- CG-HOI: Contact-Guided 3D Human-Object Interaction Generation (CVPR 2024 · similarity 0.91)
- Real2Sim in HOI: Toward Physically Plausible HOI Reconstruction from Monocular Videos (NeurIPS 2026 · similarity 0.91)
- InterDreamer: Zero-Shot Text to 3D Dynamic Human-Object Interaction (NeurIPS 2024 · similarity 0.90)
- Scaling Up Dynamic Human-Scene Interaction Modeling (CVPR 2024 · similarity 0.89)
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.
