EUVIMEDEUVIMED entdecken
Crossref

Geometry‐Aware Deep Learning for Automated Anatomical Landmark Detection on 3D Femoral Surfaces

Jianheng Zhang, Jinghao Xu, Junqing Wang, Yong Nie, Zhongmin Jin, Jinwu Wang, Junyan Li

Concurrency and Computation: Practice and Experience · 2026-07 · 38 (14)

DOI: 10.1002/cpe.70859 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Concurrency and Computation: Practice and Experience, online erschienen am 12.07.2026. Verfasst von Jianheng Zhang und 6 weiteren Autor:innen. Die Arbeit nennt 26 Literaturangaben. Gefördert von: National Natural Science Foundation of China, National Key Research and Development Program of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate localization of femoral anatomical landmarks is essential for orthopedic surgery planning, biomechanical modeling, and musculoskeletal analysis; however, existing deep learning‐based methods that directly regress landmark coordinates often suffer from limited robustness and interpretability under anatomical variability and sparse annotations. In this study, we propose a region‐aware femoral landmark detection framework that integrates deep learning‐based semantic segmentation with anatomy‐driven geometric localization. An enhanced PointNet++ network with a multiscale convolutional attention mechanism is used to refine grouped local feature representations for the segmentation of anatomically meaningful femoral regions, after which explicit geometric definitions are applied within each region to localize key landmarks, improving stability and interpretability across subjects. Validation against expert‐annotated ground truth demonstrates a mean localization error of 1.08 mm across 10 anatomical landmarks, with consistently bounded errors while maintaining a lightweight computational profile. By decoupling landmark detection into region segmentation and geometry‐based localization, the proposed approach achieves accurate, robust, and interpretable femoral landmark detection, making it well suited for clinical and biomechanical applications.

Förderung

  • National Natural Science Foundation of China
  • National Key Research and Development Program of China

Zitieren

APA

Zhang, J., Xu, J., Wang, J., Nie, Y., Jin, Z., Wang, J., & Li, J. (2026). Geometry‐Aware Deep Learning for Automated Anatomical Landmark Detection on 3D Femoral Surfaces. Concurrency and Computation: Practice and Experience, 38(14). https://doi.org/10.1002/cpe.70859

Vancouver

Zhang J, Xu J, Wang J, Nie Y, Jin Z, Wang J, et al. Geometry‐Aware Deep Learning for Automated Anatomical Landmark Detection on 3D Femoral Surfaces. Concurrency and Computation: Practice and Experience. 2026;38(14). doi:10.1002/cpe.70859

BibTeX
@article{Zhang2026,
  author = {Zhang, Jianheng and Xu, Jinghao and Wang, Junqing and Nie, Yong and Jin, Zhongmin and Wang, Jinwu and Li, Junyan},
  title = {Geometry‐Aware Deep Learning for Automated Anatomical Landmark Detection on 3D Femoral Surfaces},
  journal = {Concurrency and Computation: Practice and Experience},
  year = {2026},
  volume = {38},
  number = {14},
  doi = {10.1002/cpe.70859}
}
Kostenlos starten →