Research map / 3D vision / Face and shape
CAD-Deform: Deformable Fitting of CAD Models to 3D Scans
Vladislav Ishimtsev, Alexey Bokhovkin, Alexey Artemov, Savva Ignatyev, Matthias Niessner, Denis Zorin, Evgeny Burnaev
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 smoothness term to deformation energy (ablation) | CAD-Deform · Scan2CAD GT alignments, class avg. | accuracy (%): 90.6 -> 91.7 | Improved |
| add sharp feature term to deformation energy (ablation) | CAD-Deform · Scan2CAD GT alignments, class avg. | accuracy (%): 90.3 -> 91.7 | Improved |
| add sharp feature term (ablation) | CAD-Deform · Scan2CAD E2E alignments, class avg. | accuracy (%): 90.3 vs 90.6 | Hurt |
| add smoothness term (ablation) | CAD-Deform · Scan2CAD S2C alignments, class avg. | accuracy (%): 90.0 vs 89.8 | Hurt |
| add second-stage point-to-point optimization after nearest-neighbor stage (ablation) | CAD-Deform · Scan2CAD S2C alignments, class avg. | accuracy (%): 84.3 -> 88.3 | Improved |
| apply CAD-Deform to ground-truth alignments (ablation) | Ground-truth alignment · Scan2CAD GT alignments, instance avg. | accuracy (%): 90.6 -> 93.1 | Improved |
| apply CAD-Deform to S2C predicted alignments (ablation) | Scan2CAD alignment · Scan2CAD S2C alignments, class avg. | accuracy (%): 83.7 -> 89.8 | Improved |
| use finer PartNet part labels (level 4 vs object level) (ablation) | CAD-Deform · Scan2CAD GT alignments, class avg. | accuracy (%): 89.25 -> 91.65 | Improved |
| use CAD-Deform instead of Harmonic deformation | mesh deformation · Scan2CAD deformation quality | DAME (lower better): 65.1 -> 20.5 | Improved |
| use CAD-Deform instead of Harmonic deformation | mesh deformation · Scan2CAD fitting, class avg. | accuracy (%): 96.2 -> 91.7 | Hurt |
| use CAD-Deform instead of ARAP | mesh deformation · Scan2CAD user study | user score: 4.0 -> 7.7 | Improved |
Closest papers by meaning
- MiCADangelo: Fine-Grained Reconstruction of Constrained CAD Models from 3D Scans (NeurIPS 2025 · similarity 0.85)
- CADCrafter: Generating Computer-Aided Design Models from Unconstrained Images (CVPR 2025 · similarity 0.85)
- KP-RED: Exploiting Semantic Keypoints for Joint 3D Shape Retrieval and Deformation (CVPR 2024 · similarity 0.85)
- From 2D CAD Drawings to 3D Parametric Models: A Vision-Language Approach (AAAI 2025 · similarity 0.84)
- CAD-SIGNet: CAD Language Inference from Point Clouds using Layer-wise Sketch Instance Guided Attention (CVPR 2024 · similarity 0.84)
- cadrille: Multi-modal CAD Reconstruction with Reinforcement Learning (ICLR 2026 · similarity 0.84)
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.
