EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

GeoRefine-Net: A Geometry-Consistent Boundary Refinement Network for Industrial-Style Point Cloud Edge Detection

Yufeng Li, Kejun Wang, Xin Su, Pankoo Kim, Hyo-jai Lee

Mathematics · 20.07.2026 · 14 (14) · 2632

DOI: 10.3390/math14142632 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Mathematics, online erschienen am 20.07.2026. Verfasst von Yufeng Li und 4 weiteren Autor:innen. Die Arbeit nennt 32 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Chosun University.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Industrial-style 3D point clouds frequently exhibit intricate geometric structures such as abrupt curvature changes, multi-surface junctions, and fine surface seams. Extracting fine boundary cues from these data remains challenging due to three intertwined factors: cross-boundary neighborhood contamination caused by purely spatial neighbor selection, orientation-sensitive boundary contrast in arbitrary poses, and over-smoothing from uniform feature aggregation. To address these issues, we present GeoRefine-Net, a geometry-consistent boundary refinement framework that progressively refines boundary predictions through three dedicated mechanisms. The geometry-consistent graph construction (GCG) module incorporates multi-order geometric statistics and a learnable boundary prior to actively discourage erroneous cross-surface neighbor associations during graph construction. The Rotation-aware Boundary Contrast Encoding (RBC) module captures local directional disparities using a composite descriptor incorporating normal cross-product interaction, angular divergence, curvature discrepancy, and tangent-plane projection, thereby providing rotation-consistency-enhanced boundary cues across varying object orientations. The Confidence-Guided Boundary Refinement (CBR) module introduces a gating mechanism that adaptively amplifies or suppresses neighbor contributions based on learned consistency scores, producing sharper boundary activation while suppressing noisy responses. The main novelty of GeoRefine-Net lies in the coupled design of geometry-consistent graph construction, rotation-consistency-enhanced boundary contrast encoding, and confidence-guided aggregation, rather than in any single handcrafted descriptor or isolated network component. Additional evaluations using SymBLE, PR-AUC, directed-versus-symmetric GCG, normal and curvature perturbations, normal-free variants, coefficient sensitivity analysis, and inference-time profiling further characterize the robustness and limitations of the proposed framework. Comprehensive evaluations on the ABC CAD dataset demonstrate that GeoRefine-Net outperforms the strongest directly compared baseline by 2.8 percentage points in the F1 score.

Förderung

  • Chosun University

Zitieren

APA

Li, Y., Wang, K., Su, X., Kim, P., & Lee, H. (2026). GeoRefine-Net: A Geometry-Consistent Boundary Refinement Network for Industrial-Style Point Cloud Edge Detection. Mathematics, 14(14), 2632. https://doi.org/10.3390/math14142632

Vancouver

Li Y, Wang K, Su X, Kim P, Lee H. GeoRefine-Net: A Geometry-Consistent Boundary Refinement Network for Industrial-Style Point Cloud Edge Detection. Mathematics. 2026;14(14):2632. doi:10.3390/math14142632

BibTeX
@article{Li2026,
  author = {Li, Yufeng and Wang, Kejun and Su, Xin and Kim, Pankoo and Lee, Hyo-jai},
  title = {GeoRefine-Net: A Geometry-Consistent Boundary Refinement Network for Industrial-Style Point Cloud Edge Detection},
  journal = {Mathematics},
  year = {2026},
  volume = {14},
  number = {14},
  pages = {2632},
  doi = {10.3390/math14142632}
}
Kostenlos starten →