EUVIMEDEUVIMED entdecken
Crossref

Application and optimization of the U-Net++ model for cerebral artery segmentation based on computed tomographic angiography images

Hajin Kim, Kang-Hyeon Seo, Kyuseok Kim, Jina Shim, Youngjin Lee

European Journal of Radiology · 2025-07 · 188 · 112137

DOI: 10.1016/j.ejrad.2025.112137 ↗

Kurz gefasst

Fachartikel in der Zeitschrift European Journal of Radiology, 2025-07. Verfasst von Hajin Kim und 4 weiteren Autor:innen. Beteiligte Einrichtung: Department of Health Science, General Graduate School of Gachon University, 191, Hambakmoero, Yeonsu-gu, Incheon 21936, Republic of Korea. Hauptthemen laut MEDLINE: Computed Tomography Angiography, Deep Learning, Radiographic Image Interpretation, Computer-Assisted, Cerebral Arteries. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 21 Literaturangaben. Gefördert von: National Research Foundation of Korea.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate segmentation of cerebral arteries on computed tomography angiography (CTA) images is essential for the diagnosis and management of cerebrovascular diseases, including ischemic stroke. This study implemented a deep learning-based U-Net++ model for cerebral artery segmentation in CTA images, focusing on optimizing pruning levels by analyzing the trade-off between segmentation performance and computational cost. Dual-energy CTA and direct subtraction CTA datasets were utilized to segment the internal carotid and vertebral arteries in close proximity to the bone. We implemented four pruning levels (L1-L4) in the U-Net++ model and evaluated the segmentation performance using accuracy, intersection over union, F1-score, boundary F1-score, and Hausdorff distance. Statistical analyses were conducted to assess the significance of segmentation performance differences across pruning levels. In addition, we measured training and inference times to evaluate the trade-off between segmentation performance and computational efficiency. Applying deep supervision improved segmentation performance across all factors. While the L4 pruning level achieved the highest segmentation performance, L3 significantly reduced training and inference times (by an average of 51.56 % and 22.62 %, respectively), while incurring only a small decrease in segmentation performance (7.08 %) compared to L4. These results suggest that L3 achieves an optimal balance between performance and computational cost. This study demonstrates that pruning levels in U-Net++ models can be optimized to reduce computational cost while maintaining effective segmentation performance. By simplifying deep learning models, this approach can improve the efficiency of cerebrovascular segmentation, contributing to faster and more accurate diagnoses in clinical settings.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Computed Tomography AngiographyDeep LearningRadiographic Image Interpretation, Computer-AssistedCerebral ArteriesCerebral AngiographyCerebrovascular DisordersHumansReproducibility of Results

Stichwörter der Autor:innen

Cerebral artery segmentationComputed tomographic angiographyDeep learningOptimization of U-Net++ modelQuantitative evaluation

Autor:innen und Institutionen

  • Department of Health Science, General Graduate School of Gachon University, 191, Hambakmoero, Yeonsu-gu, Incheon 21936, Republic of Korea.
  • Institute of Human Convergence Health Science, Gachon University, 191, Hambakmoero, Yeonsu-gu, Incheon 21936, Republic of Korea.
  • Department of Radiotechnology, Wonkwang Health Science University, 514, Iksan-daero, Iksan-si, Jeonbuk-do 54538, Republic of Korea.
  • Department of Radiological Science, Gachon University, 191, Hambakmoero, Yeonsu-gu, Incheon 21936, Republic of Korea. Electronic address: yj20@gachon.ac.kr.

Förderung

  • National Research Foundation of Korea

Zitieren

APA

Kim, H., Seo, K. H., Kim, K., Shim, J., & Lee, Y. (2025). Application and optimization of the U-Net++ model for cerebral artery segmentation based on computed tomographic angiography images. European Journal of Radiology, 188, 112137. https://doi.org/10.1016/j.ejrad.2025.112137

Vancouver

Kim H, Seo KH, Kim K, Shim J, Lee Y. Application and optimization of the U-Net++ model for cerebral artery segmentation based on computed tomographic angiography images. European Journal of Radiology. 2025;188:112137. doi:10.1016/j.ejrad.2025.112137

BibTeX
@article{Kim2025,
  author = {Kim, Hajin and Seo, Kang-Hyeon and Kim, Kyuseok and Shim, Jina and Lee, Youngjin},
  title = {Application and optimization of the U-Net++ model for cerebral artery segmentation based on computed tomographic angiography images},
  journal = {European Journal of Radiology},
  year = {2025},
  volume = {188},
  pages = {112137},
  doi = {10.1016/j.ejrad.2025.112137}
}
Kostenlos starten →