Research map / 3D vision / 3D generation
PARSE: Part-Aware Relational Spatial Modeling
Yinuo Bai, Peijun Xu, Kuixiang Shao, Yuyang Jiao, Jingxuan Zhang, Kaixin Yao, Jiayuan Gu, Jingyi Yu
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 |
|---|---|---|---|
| fine-tune Qwen3-VL on PARSE-10K | Qwen3-VL · Visual Relation MCQ | accuracy: 86.2 -> 97.4 | Improved |
| fine-tune Qwen3-VL on PARSE-10K | Qwen3-VL · Part-level Contact MCQ | accuracy: 60.4 -> 86.2 | Improved |
| fine-tune Qwen3-VL on PARSE-10K | Qwen3-VL · Scene Graph Generation (with BBox matching) | F1: 33.2 -> 76.6 | Improved |
| fine-tune Qwen3-VL on PARSE-10K | Qwen3-VL · Scene Graph Generation (no BBox matching) | F1: 37.9 -> 78.2 | Improved |
| train scene generator on PARSE-10K instead of 3D-FRONT (no PAG) | InstructScene · User study | Complexity: 7.5% -> 45.0% | Improved |
| train scene generator on PARSE-10K instead of 3D-FRONT (no PAG) | InstructScene · User study | Realism: 33.8% -> 27.5% | Hurt |
| add PAG conditioning (ablation) | Scene generator trained on PARSE-10K · User study | Contact Fidelity: 26.3% -> 45.0% | Improved |
| add PAG conditioning (ablation) | Scene generator trained on PARSE-10K · User study | Realism: 27.5% -> 38.8% | Improved |
| add PAG conditioning (ablation) | Scene generator trained on PARSE-10K · User study | Complexity: 45.0% -> 47.5% | Improved |
Closest papers by meaning
- Thinking in Structures: Evaluating Spatial Intelligence in Constraint-Governed Spaces (ICML 2026 · similarity 0.89)
- Surprise3D: A Dataset for Spatial Understanding and Reasoning in Complex 3D Scenes (NeurIPS 2025 · similarity 0.87)
- Direct Numerical Layout Generation for 3D Indoor Scene Synthesis via Spatial Reasoning (NeurIPS 2025 · similarity 0.87)
- LayoutVLM: Differentiable Optimization of 3D Layout via Vision-Language Models (CVPR 2025 · similarity 0.87)
- SceneFun3D: Fine-Grained Functionality and Affordance Understanding in 3D Scenes (CVPR 2024 · similarity 0.87)
- SpaceVista: All-Scale Visual Spatial Reasoning from mm to km (ICML 2026 · 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.
