Research map / 3D vision / Face and shape
Scan2CAD: Learning CAD Model Alignment in RGB-D Scans
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 learned scan-to-CAD correspondence and alignment approach | Scan2CAD alignment benchmark · Scan2CAD alignment benchmark | accuracy: 10.29 -> 31.68 | Helped |
| add compatibility prediction (ablation) | Scan2CAD correspondence prediction · Scan2CAD correspondence prediction | F1-score: 50.85 -> 60.43 | Helped |
| add compatibility prediction (ablation) | Scan2CAD alignment model · Scan2CAD alignment benchmark | accuracy: 11.48 -> 31.68 | Helped |
| add symmetry awareness (ablation) | Scan2CAD correspondence prediction · Scan2CAD correspondence prediction | F1-score: 47.64 -> 60.43 | Helped |
| add symmetry awareness (ablation) | Scan2CAD alignment model · Scan2CAD alignment benchmark | accuracy: 26.75 -> 31.68 | Helped |
| add scale prediction (ablation) | Scan2CAD correspondence prediction · Scan2CAD correspondence prediction | F1-score: 48.01 -> 50.85 | Helped |
| add scale prediction (ablation) | Scan2CAD alignment model · Scan2CAD alignment benchmark | accuracy: 29.42 -> 31.68 | Helped |
| add encoder pre-training with 1/3 parts fixed (ablation) | Scan2CAD correspondence prediction · Scan2CAD correspondence prediction | F1-score: 60.43 -> 63.94 | Helped |
| add encoder pre-training with 3/3 parts fixed (ablation) | Scan2CAD correspondence prediction · Scan2CAD correspondence prediction | F1-score: 60.43 -> 58.61 | Hurt |
| add encoder pre-training with 1/3 parts fixed (ablation) | Scan2CAD alignment model · Scan2CAD alignment benchmark | accuracy: 31.68 -> 31.22 | No clear effect |
| add encoder pre-training with 3/3 parts fixed (ablation) | Scan2CAD alignment model · Scan2CAD alignment benchmark | accuracy: 31.68 -> 29.27 | Hurt |
Closest papers by meaning
- Point2CAD: Reverse Engineering CAD Models from 3D Point Clouds (CVPR 2024 · similarity 0.87)
- cadrille: Multi-modal CAD Reconstruction with Reinforcement Learning (ICLR 2026 · similarity 0.87)
- From 2D CAD Drawings to 3D Parametric Models: A Vision-Language Approach (AAAI 2025 · similarity 0.86)
- MiCADangelo: Fine-Grained Reconstruction of Constrained CAD Models from 3D Scans (NeurIPS 2025 · similarity 0.86)
- ReCAD: Reinforcement Learning Enhanced Parametric CAD Model Generation with Vision-Language Models (AAAI 2026 · similarity 0.85)
- CADCrafter: Generating Computer-Aided Design Models from Unconstrained Images (CVPR 2025 · similarity 0.85)
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.