POPCORN: Progressive Pseudo-labeling with Consistency Regularization and Neighboring - Archive ouverte HAL Access content directly
Conference Papers Year : 2021

POPCORN: Progressive Pseudo-labeling with Consistency Regularization and Neighboring

Abstract

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.
Fichier principal
Vignette du fichier
POPCORN_Arxiv_version.pdf (1.8 Mo) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03365185 , version 1 (14-10-2021)

Identifiers

Cite

Reda Abdellah Kamraoui, Vinh-Thong Ta, Nicolas Papadakis, Fanny Compaire, José V Manjon, et al.. POPCORN: Progressive Pseudo-labeling with Consistency Regularization and Neighboring. International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI'21), Sep 2021, Strasbourg (virtual), France. pp.373-382, ⟨10.1007/978-3-030-87196-3_35⟩. ⟨hal-03365185⟩

Collections

CNRS ANR
115 View
144 Download

Altmetric

Share

Gmail Facebook X LinkedIn More