• Mon espace de travail
  • Aide IRIS
  • Par Publication Par Personne Par Unité
    • English
    • Français
  • Se connecter
Logo du site

IRIS | Système d’Information de la Recherche Institutionnelle

  • Accueil
  • Personnes
  • Publications
  • Unités
  • Périodiques
UNIL
  • English
  • Français
Se connecter
IRIS
  • Accueil
  • Personnes
  • Publications
  • Unités
  • Périodiques
  • Mon espace de travail
  • Aide IRIS

Parcourir IRIS

  • Par Publication
  • Par Personne
  • Par Unité
  1. Accueil
  2. IRIS
  3. Publication
  4. Label-efficient deep semantic segmentation of intracranial hemorrhages in CT-scans.
 
  • Détails
Titre

Label-efficient deep semantic segmentation of intracranial hemorrhages in CT-scans.

Type
article
Institution
UNIL/CHUV/Unisanté + institutions partenaires
Périodique
Frontiers in Neuroimaging  
Auteur(s)
Spahr, A.
Auteure/Auteur
Ståhle, J.
Auteure/Auteur
Wang, C.
Auteure/Auteur
Kaijser, M.
Auteure/Auteur
Liens vers les unités
Recherche en neurosciences  
ISSN
2813-1193
Statut éditorial
Publié
Date de publication
2023-07
Volume
2
Première page
1157565
Peer-reviewed
Oui
Langue
anglais
Notes
Publication types: Journal Article
Publication Status: epublish
Résumé
Intracranial hemorrhage (ICH) is a common finding in traumatic brain injury (TBI) and computed tomography (CT) is considered the gold standard for diagnosis. Automated detection of ICH provides clinical value in diagnostics and in the ability to feed robust quantification measures into future prediction models. Several studies have explored ICH detection and segmentation but the research process is somewhat hindered due to a lack of open large and labeled datasets, making validation and comparison almost impossible. The complexity of the task is further challenged by the heterogeneity of ICH patterns, requiring a large number of labeled data to train robust and reliable models. Consequently, due to the labeling cost, there is a need for label-efficient algorithms that can exploit easily available unlabeled or weakly-labeled data. Our aims for this study were to evaluate whether transfer learning can improve ICH segmentation performance and to compare a variety of transfer learning approaches that harness unlabeled and weakly-labeled data. Three self-supervised and three weakly-supervised transfer learning approaches were explored. To be used in our comparisons, we also manually labeled a dataset of 51 CT scans. We demonstrate that transfer learning improves ICH segmentation performance on both datasets. Unlike most studies on ICH segmentation our work relies exclusively on publicly available datasets, allowing for easy comparison of performances in future studies. To further promote comparison between studies, we also present a new public dataset of ICH-labeled CT scans, Seq-CQ500.
Sujets

ICH segmentation

computed tomography

computer vision

dataset

transfer learning

traumatic brain injur...

PID Serval
serval:BIB_77E94F69A0D3
DOI
10.3389/fnimg.2023.1157565
PMID
37554648
Permalien
https://iris.unil.ch/handle/iris/181437
Open Access
Oui
Date de création
2023-08-10T12:24:25.825Z
Date de création dans IRIS
2025-05-21T01:00:45Z
Fichier(s)
En cours de chargement...
Vignette d'image
Nom

37554648_BIB_77E94F69A0D3.pdf

Version du manuscrit

published

Licence

https://creativecommons.org/licenses/by/4.0

Taille

1.78 MB

Format

Adobe PDF

PID Serval

serval:BIB_77E94F69A0D3.P001

URN

urn:nbn:ch:serval-BIB_77E94F69A0D38

Somme de contrôle

(MD5):7b6aa6ca1225caf8be8c787a1f767a67

  • Copyright © 2024 UNIL
  • Informations légales