Research map / Video / Video generation
SIMS: Simulating Stylized Human-Scene Interactions with Retrieval-Augmented Script Generation
Wenjia Wang, Liang Pan, Zhiyang Dou, Jidong Mei, Zhouyingcheng Liao, Yuke Lou, Yifan Wu, Lei Yang, Jingbo Wang, Taku Komura
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 |
|---|---|---|---|
| use RASG instead of direct GPT-4 script generation (ablation) | script planner · 200 generated scripts | SBERT similarity (lower more diverse): 0.8167 -> 0.7759 | Improved |
| use RASG instead of direct GPT-4 script generation (ablation) | script planner · 200 generated scripts | generation time (s): 12.2 -> 7.32 | Improved |
| add text embedding conditioning (ablation) | SIMS policy · Sit skill | success rate (%): 89.7 -> 96.9 | Improved |
| add text embedding conditioning (ablation) | SIMS policy · Carry skill | APD: 12.41 -> 14.92 | Improved |
| add text embedding conditioning (ablation) | SIMS policy · Lie skill | success rate (%): 89.6 -> 89.7 | No clear effect |
| add egocentric heightmap (ablation) | SIMS policy · Sit skill | success rate (%): 88.7 -> 96.9 | Improved |
| add egocentric heightmap (ablation) | SIMS policy · Lie skill | success rate (%): 79.8 -> 89.7 | Improved |
| add AMASS to 100Style training data (ablation) | Walk policy · Walk skill | success rate (%): 92.6 -> 95.1 | Improved |
| add ViconStyle carry motions to AMASS (ablation) | Carry policy · Carry skill | success rate (%): 92.9 -> 96.4 | Improved |
| add COUCH data to SAMP (ablation) | Sit policy · Sit skill | success rate (%): 95.5 -> 96.9 | Improved |
| add ViconStyle to SAMP and COUCH (ablation) | Lie policy · Lie skill | success rate (%): 86.9 -> 89.7 | Improved |
| evaluate on unseen PartNet objects without retraining | SIMS Sit policy · PartNet vs 3DFront | success rate (%): 96.9 -> 98.7 | No clear effect |
Closest papers by meaning
- Sitcom-Crafter: A Plot-Driven Human Motion Generation System in 3D Scenes (ICLR 2025 · similarity 0.87)
- InterDreamer: Zero-Shot Text to 3D Dynamic Human-Object Interaction (NeurIPS 2024 · similarity 0.87)
- HSI-GPT: A General-Purpose Large Scene-Motion-Language Model for Human Scene Interaction (CVPR 2025 · similarity 0.86)
- TIMotion: Temporal and Interactive Framework for Efficient Human-Human Motion Generation (CVPR 2025 · similarity 0.86)
- SemGeoMo: Dynamic Contextual Human Motion Generation with Semantic and Geometric Guidance (CVPR 2025 · similarity 0.86)
- InterControl: Zero-shot Human Interaction Generation by Controlling Every Joint (NeurIPS 2024 · 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.
