Research map / Point clouds / Point cloud classification
CMT: A Cascade MAR with Topology Predictor for Multimodal Conditional CAD Generation
Jianyu Wu, Yizhou Wang, Xiangyu Yue, Xinzhu Ma, Jinyang Guo, Dongzhan Zhou, Wanli Ouyang, Shixiang Tang
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 cascade edges-then-surfaces network instead of a single network (ablation) | CMT · DeepCAD | COV: 65.80 -> 75.71 | Improved |
| add cascade network (ablation) | CMT · DeepCAD | Valid: 47.17 -> 70.13 | Improved |
| add cascade network (ablation) | CMT · DeepCAD | JSD: 4.71 -> 1.02 | Improved |
| increase sampling steps from 16/8 to 64/32 (ablation) | CMT · DeepCAD | Valid: 47.80 -> 70.13 | Improved |
| increase sampling steps from 32/16 to 64/32 (ablation) | CMT · DeepCAD | COV: 74.97 -> 75.71 | Improved |
| use 1/1 sampling steps (all tokens at once) (ablation) | CMT · DeepCAD | Valid: 70.13 -> 0.10 | Hurt |
| use CMT instead of BrepGen | BrepGen · ABC unconditional | COV: 57.92 -> 68.60 | Improved |
| use CMT instead of BrepGen | BrepGen · DeepCAD unconditional | Valid: 62.9 -> 70.1 | Improved |
| use CMT instead of DeepCAD for point-cloud conditioning | DeepCAD · point-cloud generation | Chamfer: 5.11 -> 0.64 | Improved |
| use CMT instead of InstantMesh for single-image conditioning | InstantMesh · image-conditioned generation | Chamfer: 6.18 -> 2.17 | Improved |
| use CMT instead of Michelangelo for text conditioning | Michelangelo · text-conditioned generation | human win rate: 20 -> 80 | Improved |
Closest papers by meaning
- CAD-GPT: Synthesising CAD Construction Sequence with Spatial Reasoning-Enhanced Multimodal LLMs (AAAI 2025 · similarity 0.89)
- cadrille: Multi-modal CAD Reconstruction with Reinforcement Learning (ICLR 2026 · similarity 0.88)
- CADCrafter: Generating Computer-Aided Design Models from Unconstrained Images (CVPR 2025 · similarity 0.88)
- Draw Step by Step: Reconstructing CAD Construction Sequences from Point Clouds via Multimodal Diffusion. (CVPR 2024 · similarity 0.87)
- FlexCAD: Unified and Versatile Controllable CAD Generation with Fine-tuned Large Language Models (ICLR 2025 · similarity 0.87)
- GeoCAD: Local Geometry-Controllable CAD Generation with Large Language Models (NeurIPS 2025 · 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.
