EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

GraphBRep : Explicit Graph Diffusion of B–Rep Topology for Efficient CAD Generation

Weilin Lai, Tie Xu, Ben Niu, Hu Wang

Journal of Computational Design and Engineering · 02.01.2026 · 13 (1) · 259-274

DOI: 10.1093/jcde/qwaf136 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Computational Design and Engineering, online erschienen am 16.12.2025. Verfasst von Weilin Lai und 3 weiteren Autor:innen. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0). Gefördert von: National Natural Science Foundation of China, Peacock Program for Overseas High-Level Talents Introduction of Shenzhen City.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Direct B-Rep generation is increasingly important in Computer-Aided Design (CAD) workflows, eliminating costly modeling sequence data and enabling the creation of complex features. A key challenge is modeling the joint distribution of the misaligned geometry and topology. Existing methods tend to implicitly embed topology into the geometric features of edges. Although this integration ensures feature alignment, it also causes edge geometry to carry more redundant structural information compared to the original B-Rep, leading to significantly higher computational cost. For efficient generation, GraphBRep, a B-Rep generation model that explicitly represents and learns compact topology, is proposed. Based on the original B-Rep structure, an undirected weighted graph is constructed to represent the surface topology. A graph diffusion model is employed to learn topology conditioned on surface features, serving as the basis for determining connectivity between primitive surfaces. The explicit representation ensures a compact data structure, effectively reducing computational cost during both training and inference. Experiments on two large-scale unconditional datasets and one category-conditional dataset demonstrate that the proposed method significantly reduces training and inference times (up to 30.82% and 56.3% for given datasets, respectively) while maintaining high-quality CAD generation compared with SOTA.

Förderung

  • National Natural Science Foundation of China
  • Peacock Program for Overseas High-Level Talents Introduction of Shenzhen City

Zitieren

APA

Lai, W., Xu, T., Niu, B., & Wang, H. (2026). GraphBRep : Explicit Graph Diffusion of B–Rep Topology for Efficient CAD Generation. Journal of Computational Design and Engineering, 13(1), 259-274. https://doi.org/10.1093/jcde/qwaf136

Vancouver

Lai W, Xu T, Niu B, Wang H. GraphBRep : Explicit Graph Diffusion of B–Rep Topology for Efficient CAD Generation. Journal of Computational Design and Engineering. 2026;13(1):259-274. doi:10.1093/jcde/qwaf136

BibTeX
@article{Lai2026,
  author = {Lai, Weilin and Xu, Tie and Niu, Ben and Wang, Hu},
  title = {GraphBRep : Explicit Graph Diffusion of B–Rep Topology for Efficient CAD Generation},
  journal = {Journal of Computational Design and Engineering},
  year = {2026},
  volume = {13},
  number = {1},
  pages = {259--274},
  doi = {10.1093/jcde/qwaf136}
}
Kostenlos starten →