Research map / 3D vision / Face and shape
ShapeScaffolder: Structure-Aware 3D Shape Generation from Text
Xi Tian, Yong-Liang Yang, Qi Wu
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 |
|---|---|---|---|
| replace flat decoder with hierarchical decoder (ablation) | HierD · Level 1 output | IoU: 10.38 -> 11.20 | Improved |
| replace word-level attention with graph-part attention at Level 2 (ablation) | HierD · Level 2 output | IoU: 11.66 -> 12.63 | Improved |
| add graph-part attention at Level 2 to base hierarchical decoder (ablation) | HierD · Level 2 output | CLIP: 45.36 -> 50.64 | Improved |
| apply graph attention at Level 1 instead of Level 2 (ablation) | HierD · Level 2 output | IoU: 12.63 -> 11.30 | Hurt |
| add joint refiner at Level 2 (ablation) | HierD + GraphAttn · Level 2 output | IoU: 12.63 -> 13.65 | Improved |
| add joint refiner at Level 2 (ablation) | HierD + GraphAttn · Level 2 output | FID: 16.31 -> 15.81 | Improved |
| add word-level attention at Level 2 (ablation) | HierD · Level 2 output | IoU: 11.07 -> 11.66 | Improved |
| add hierarchical bond loss structure with deeper levels (ablation) | HierD · Level 1 vs Level 2 output | IoU: 11.20 -> 11.07 | Hurt |
| structure-aware text-to-shape generation versus T2S-Implicit | ShapeScaffolder · Text2Shape dataset | Acc: 97.48 -> 97.5 | No clear effect |
Closest papers by meaning
- Towards Implicit Text-Guided 3D Shape Generation (CVPR 2022 · similarity 0.93)
- HyperSDFusion: Bridging Hierarchical Structures in Language and Geometry for Enhanced 3D Text2Shape Generation (CVPR 2024 · similarity 0.89)
- ShapeWords: Guiding Text-to-Image Synthesis with 3D Shape-Aware Prompts (CVPR 2025 · similarity 0.89)
- Shape of Thought: Progressive Object Assembly via Visual Chain-of-Thought (ICML 2026 · similarity 0.88)
- PI3D: Efficient Text-to-3D Generation with Pseudo-Image Diffusion (CVPR 2024 · similarity 0.87)
- ShapeWalk: Compositional Shape Editing Through Language-Guided Chains (CVPR 2024 · similarity 0.87)
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.
