: Scaling Multi-modal CAD Generation using Differentiable Parametric Surfaces
Muhammad Sadil Khan 1,2† Muhammad Usama 1,2 Rolandos Alexandros Potamias 3 Didier Stricker 1,2 Muhammad Zeshan Afzal 1 Jiankang Deng 3 Ismail Elezi 4 1 DFKI 2 RPTU, Kaiserslautern 3 Imperial College London 4 Huawei London Research Center · 2026
Read : Scaling Multi-modal CAD Generation using Differentiable Parametric Surfaces by Muhammad Sadil Khan 1,2† Muhammad Usama 1,2 Rolandos Alexandros Potamias 3 Didier Stricker 1,2 Muhammad Zeshan Afzal 1 Jiankang Deng 3 Ismail Elezi 4 1 DFKI 2 RPTU, Kaiserslautern 3 Imperial College London 4 Huawei London Research Center (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-05607.txt.
Description
Khan and team propose DreamCAD, a multi-modal framework that makes edge forms from point-level guidance alone, with no CAD-only annotations. Each edge form becomes a set of patches set by settings, such as Bézier faces, made through smooth tiling, for large-scale training with linked surfaces to edit. They also debut CADCap-1M, the largest CAD caption dataset, with over 1M notes made using GPT-5. DreamCAD leads on ABC and Objaverse across text, image, and point modes, with better shape trueness and passing 75% user liking.
Catalog record
| CREATOR | Muhammad Sadil Khan 1,2† Muhammad Usama 1,2 Rolandos Alexandros Potamias 3 Didier Stricker 1,2 Muhammad Zeshan Afzal 1 Jiankang Deng 3 Ismail Elezi 4 1 DFKI 2 RPTU, Kaiserslautern 3 Imperial College London 4 Huawei London Research Center |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 37842 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.05607