EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Region-based evidential deep learning to quantify uncertainty and improve robustness of brain tumor segmentation

Hao Li, Yang Nan, Javier Del Ser, Guang Yang iD

Neural Computing and Applications · 2023-10 · 35 (30) · 22071-22085

DOI: 10.1007/s00521-022-08016-4 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Neural Computing and Applications, online erschienen am 17.11.2022. Verfasst von Hao Li und 3 weiteren Autor:innen. Beteiligte Einrichtung: National Heart and Lung Institute, Faculty of Medicine, Imperial College London, London, UK. Die Arbeit wurde bisher 63-mal zitiert (Crossref) und nennt 36 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Medical Research Council, British Heart Foundation, ERC IMI.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Despite recent advances in the accuracy of brain tumor segmentation, the results still suffer from low reliability and robustness. Uncertainty estimation is an efficient solution to this problem, as it provides a measure of confidence in the segmentation results. The current uncertainty estimation methods based on quantile regression, Bayesian neural network, ensemble, and Monte Carlo dropout are limited by their high computational cost and inconsistency. In order to overcome these challenges, Evidential Deep Learning (EDL) was developed in recent work but primarily for natural image classification and showed inferior segmentation results. In this paper, we proposed a region-based EDL segmentation framework that can generate reliable uncertainty maps and accurate segmentation results, which is robust to noise and image corruption. We used the Theory of Evidence to interpret the output of a neural network as evidence values gathered from input features. Following Subjective Logic, evidence was parameterized as a Dirichlet distribution, and predicted probabilities were treated as subjective opinions. To evaluate the performance of our model on segmentation and uncertainty estimation, we conducted quantitative and qualitative experiments on the BraTS 2020 dataset. The results demonstrated the top performance of the proposed method in quantifying segmentation uncertainty and robustly segmenting tumors. Furthermore, our proposed new framework maintained the advantages of low computational cost and easy implementation and showed the potential for clinical application.

Schlagwörter

Stichwörter der Autor:innen

Brain tumor segmentationEvidential deep learningRobustnessUncertainty quantification

Autor:innen und Institutionen

  • National Heart and Lung Institute, Faculty of Medicine, Imperial College London, London, UK.
  • Department of Bioengineering, Faculty of Engineering, Imperial College London, London, UK.
  • TECNALIA, Basque Research and Technology Alliance (BRTA), Derio, Spain.
  • University of the Basque Country (UPV/EHU), Bilbao, Spain.
  • Royal Brompton Hospital, London, UK.

ORCID-Profile: Guang Yang

Förderung

  • Medical Research Council
  • British Heart Foundation
  • ERC IMI
  • Horizon 2020 Framework Programme
  • Royal Society
  • Imperial College Undergraduate Research Opportunities Programme
  • NVIDIA Academic Hardware Grant Program
  • Boehringer Ingelheim

Zitieren

APA

Li, H., Nan, Y., Del Ser, J., & Yang, G. (2023). Region-based evidential deep learning to quantify uncertainty and improve robustness of brain tumor segmentation. Neural Computing and Applications, 35(30), 22071-22085. https://doi.org/10.1007/s00521-022-08016-4

Vancouver

Li H, Nan Y, Del Ser J, Yang G. Region-based evidential deep learning to quantify uncertainty and improve robustness of brain tumor segmentation. Neural Computing and Applications. 2023;35(30):22071-22085. doi:10.1007/s00521-022-08016-4

BibTeX
@article{Li2023,
  author = {Li, Hao and Nan, Yang and Del Ser, Javier and Yang, Guang},
  title = {Region-based evidential deep learning to quantify uncertainty and improve robustness of brain tumor segmentation},
  journal = {Neural Computing and Applications},
  year = {2023},
  volume = {35},
  number = {30},
  pages = {22071--22085},
  doi = {10.1007/s00521-022-08016-4}
}
Kostenlos starten →