Research map / 3D vision / 3D generation
Narrator: Towards Natural Control of Human-Scene Interaction Generation via Relationship Reasoning
Haibiao Xuan, Xiongzheng Li, Jinsong Zhang, Hongwen Zhang, Yebin Liu, Kun Li
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 the JGLSG scene-graph mechanism instead of BERT text encoding (ablation) | Narrator · PROX | non-collision score: 0.95 -> 0.98 | Improved |
| add the JGLSG scene-graph mechanism instead of BERT text encoding (ablation) | Narrator · PROX | diversity entropy: 3.82 -> 4.16 | Improved |
| add the part-level action (PLA) mechanism (ablation) | Narrator · PROX | non-collision score: 0.89 -> 0.98 | Improved |
| add the part-level action (PLA) mechanism (ablation) | Narrator · PROX | diversity entropy: 3.73 -> 4.16 | Improved |
| add the IBS loss (ablation) | Narrator · PROX | contact score: 0.90 -> 0.94 | Improved |
| add the IBS loss (ablation) | Narrator · PROX | non-collision score: 0.94 -> 0.98 | Improved |
| Narrator versus COINS-Text baseline | text-conditioned HSI generation · PROX | perceptual accuracy: 3.48 -> 4.02 | Improved |
| Narrator versus POSA-Text on collision | text-conditioned HSI generation · PROX | non-collision score: 0.96 -> 0.98 | Improved |
Closest papers by meaning
- HSI-GPT: A General-Purpose Large Scene-Motion-Language Model for Human Scene Interaction (CVPR 2025 · similarity 0.89)
- Learning to Generate Human-Human-Object Interactions from Textual Descriptions (NeurIPS 2025 · similarity 0.89)
- InteractAnything: Zero-shot Human Object Interaction Synthesis via LLM Feedback and Object Affordance Parsing (CVPR 2025 · similarity 0.88)
- Unified Human-Scene Interaction via Prompted Chain-of-Contacts (ICLR 2024 · similarity 0.88)
- CG-HOI: Contact-Guided 3D Human-Object Interaction Generation (CVPR 2024 · similarity 0.88)
- InterDreamer: Zero-Shot Text to 3D Dynamic Human-Object Interaction (NeurIPS 2024 · similarity 0.88)
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.
