KAMRAOUI, Reda Abdellah; TA, Vinh-Thong; PAPADAKIS, Nicolas; COMPAIRE, Fanny; MANJON, José V; COUPÉ, Pierrick

doi:10.1007/978-3-030-87196-3_35

Métadonnées

Afficher la notice complète

Licence d’utilisation du document

KAMRAOUI, Reda Abdellah
Laboratoire Bordelais de Recherche en Informatique [LaBRI]

TA, Vinh-Thong
Laboratoire Bordelais de Recherche en Informatique [LaBRI]

PAPADAKIS, Nicolas
Institut de Mathématiques de Bordeaux [IMB]

Langue

Communication dans un congrès

Ce document a été publié dans

International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI'21), 2021-09-27, Strasbourg (virtual). 2021, vol. 12902, p. 373-382

Date de soutenance

2021

Résumé en anglais

Semi-supervised learning (SSL) uses unlabeled data to compensate for the scarcity of annotated images and the lack of method generalization to unseen domains, two usual problems in medical segmentation tasks. In this work, we propose POPCORN, a novel method combining consistency regularization and pseudo-labeling designed for image segmentation. The proposed framework uses high-level regularization to constrain our segmentation model to use similar latent features for images with similar segmentations. POPCORN estimates a proximity graph to select data from easiest ones to more difficult ones, in order to ensure accurate pseudo-labeling and to limit confirmation bias. Applied to multiple sclerosis lesion segmentation, our method demonstrates competitive results compared to other state-of-the-art SSL strategies.< Réduire

Mots clés en anglais

MS lesion segmentation

Consistency regularization

Pseudo-labeling

Semi-supervised Learning

URI

https://oskar-bordeaux.fr/handle/20.500.12278/191439

DOI

http://dx.doi.org/10.1007/978-3-030-87196-3_35

Project ANR

Apprentissage profond pour la volumétrie cérébrale : vers le BigData en neuroscience - ANR-18-CE45-0013

Origine

Importé de hal

Unités de recherche

Institut de Mathématiques de Bordeaux (IMB) - UMR 5251