Research map / 3D vision / Depth and stereo
Vid2CAD: CAD Model Alignment Using Multi-View Constraints From Videos
Kevis-Kokitsi Maninis, Stefan Popov, Matthias Nießner, Vittorio Ferrari
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 multi-view temporal integration to single-frame Mask2CAD | Mask2CAD b5 · Scan2CAD val | class avg accuracy: 11.6 -> 30.7 | Improved |
| add predicted scale and derived depth instead of class-average values | Mask2CAD · Scan2CAD val | class avg accuracy: 2.5 -> 12.1 | Improved |
| use ground-truth association instead of thresholding (ablation) | Mask2CAD · Scan2CAD val | class avg accuracy: 11.6 -> 12.1 | No clear effect |
| use multi-view reprojection scale instead of predicted scale only (ablation) | Vid2CAD · Scan2CAD val (gt association) | class avg accuracy: 30.2 -> 33.5 | Improved |
| combine multi-view and predicted scale (ablation) | Vid2CAD · Scan2CAD val (gt association) | class avg accuracy: 33.5 -> 37.4 | Improved |
| add rotation symmetry term (ablation) | Vid2CAD · Scan2CAD val (gt association) | class avg accuracy: 37.4 -> 37.6 | No clear effect |
| use automatic tracker instead of ground-truth association (ablation) | Vid2CAD · Scan2CAD val | class avg accuracy: 37.6 -> 30.7 | Hurt |
| add temporal integration to single-frame baseline with gt association | b4 · Scan2CAD val (gt association) | class avg accuracy: 12.1 -> 37.6 | Improved |
| use Vid2CAD instead of MVS + RGB-D alignment | RGB-D CAD alignment on DVMVS · Scan2CAD val | class avg accuracy: 18.8 -> 30.7 | Improved |
| use Vid2CAD instead of ODAM at strict IoU | ODAM · ScanNet | F1 @ IoU>0.7: 3.6 -> 10.5 | Improved |
| use Vid2CAD instead of ODAM at loose IoU | ODAM · ScanNet | F1 @ IoU>0.25: 61.5 -> 56.3 | Hurt |
| temporal integration for CAD retrieval | Mask2CAD b5 · Scan2CAD aligned objects | mIoU: 73.1 -> 85 | Improved |
Later papers that compared with this method
2 later accepted papers report Vid2CAD as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- FastCAD: Real-Time CAD Retrieval and Alignment from Scans and Videos (ECCV 2024 · better on 4 of 4 shared results)
- RayTran: 3D Pose Estimation and Shape Reconstruction of Multiple Objects from Videos with Ray-Traced Transformers (ECCV 2022 · better on 4 of 4 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- VideoCAD: A Dataset and Model for Learning Long‑Horizon 3D CAD UI Interactions from Video (NeurIPS 2025 · similarity 0.85)
- From 2D CAD Drawings to 3D Parametric Models: A Vision-Language Approach (AAAI 2025 · similarity 0.85)
- Point2CAD: Reverse Engineering CAD Models from 3D Point Clouds (CVPR 2024 · similarity 0.85)
- cadrille: Multi-modal CAD Reconstruction with Reinforcement Learning (ICLR 2026 · similarity 0.85)
- CADCrafter: Generating Computer-Aided Design Models from Unconstrained Images (CVPR 2025 · similarity 0.84)
- MiCADangelo: Fine-Grained Reconstruction of Constrained CAD Models from 3D Scans (NeurIPS 2025 · 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.
