Research map / 3D vision / 3D generation
ViHOI: Human-Object Interaction Synthesis with Visual Priors
Songjin Cai, Linjie Zhong, Ling Guo, Changxing Ding
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 ViHOI visual priors | CHOIS · FullBodyManipulation | MPJPE (cm): 15.43 -> 14.97 | Improved |
| add ViHOI visual priors | ROG · FullBodyManipulation | MPJPE (cm): 22.59 -> 22.61 | No clear effect |
| add ViHOI visual priors | MDM · FullBodyManipulation | MPJPE (cm): 23.11 -> 23.12 | No clear effect |
| add ViHOI visual priors | MDM · FullBodyManipulation | FID: 1.56 -> 1.18 | Improved |
| add ViHOI visual priors | ROG · FullBodyManipulation | FID: 1.54 -> 0.68 | Improved |
| add ViHOI visual priors | CHOIS · FullBodyManipulation (unseen objects) | FID: 4.99 -> 2.02 | Improved |
| add ViHOI visual priors | CHOIS · BEHAVE | MPJPE (cm): 18.51 -> 17.09 | Improved |
| replace Q-Former adaptor with average pooling + MLP (ablation) | ViHOI · FullBodyManipulation | FID: 0.68 -> 26.03 | Hurt |
| use VLM LLM text prior instead of CLIP text encoder (ablation) | ViHOI · FullBodyManipulation | MPJPE (cm): 17.57 -> 14.97 | Improved |
| use decoupled layers (visual 3, text 12) instead of other combos (ablation) | ViHOI · FullBodyManipulation | R-score: 0.61-0.75 -> 0.79 | Improved |
| add visual priors to textual prior (T12-only) (ablation) | ViHOI · FullBodyManipulation | MPJPE (cm): 17.49 -> 14.97 | Improved |
Closest papers by meaning
- CG-HOI: Contact-Guided 3D Human-Object Interaction Generation (CVPR 2024 · similarity 0.90)
- InterAct: Advancing Large-Scale Versatile 3D Human-Object Interaction Generation (CVPR 2025 · similarity 0.90)
- Learning to Generate Human-Human-Object Interactions from Textual Descriptions (NeurIPS 2025 · similarity 0.90)
- SViMo: Synchronized Diffusion for Video and Motion Generation in Hand-object Interaction Scenarios (NeurIPS 2025 · similarity 0.90)
- GenZI: Zero-Shot 3D Human-Scene Interaction Generation (CVPR 2024 · similarity 0.89)
- InteractAnything: Zero-shot Human Object Interaction Synthesis via LLM Feedback and Object Affordance Parsing (CVPR 2025 · 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.
