English

Text2CAD: Generating Sequential CAD Models from Beginner-to-Expert Level Text Prompts

Computer Vision and Pattern Recognition 2024-09-26 v1 Graphics

Abstract

Prototyping complex computer-aided design (CAD) models in modern softwares can be very time-consuming. This is due to the lack of intelligent systems that can quickly generate simpler intermediate parts. We propose Text2CAD, the first AI framework for generating text-to-parametric CAD models using designer-friendly instructions for all skill levels. Furthermore, we introduce a data annotation pipeline for generating text prompts based on natural language instructions for the DeepCAD dataset using Mistral and LLaVA-NeXT. The dataset contains 170\sim170K models and 660\sim660K text annotations, from abstract CAD descriptions (e.g., generate two concentric cylinders) to detailed specifications (e.g., draw two circles with center (x,y)(x,y) and radius r1r_{1}, r2r_{2}, and extrude along the normal by dd...). Within the Text2CAD framework, we propose an end-to-end transformer-based auto-regressive network to generate parametric CAD models from input texts. We evaluate the performance of our model through a mixture of metrics, including visual quality, parametric precision, and geometrical accuracy. Our proposed framework shows great potential in AI-aided design applications. Our source code and annotations will be publicly available.

Keywords

Cite

@article{arxiv.2409.17106,
  title  = {Text2CAD: Generating Sequential CAD Models from Beginner-to-Expert Level Text Prompts},
  author = {Mohammad Sadil Khan and Sankalp Sinha and Talha Uddin Sheikh and Didier Stricker and Sk Aziz Ali and Muhammad Zeshan Afzal},
  journal= {arXiv preprint arXiv:2409.17106},
  year   = {2024}
}

Comments

Accepted in NeurIPS 2024 (Spotlight)

R2 v1 2026-06-28T18:56:55.400Z