Synthetic tumor insertion using one-shot generative learning for cross-modal image segmentation - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2021

Synthetic tumor insertion using one-shot generative learning for cross-modal image segmentation

Résumé

Unpaired cross-modal translation with cyclic loss is being increasingly used for a large variety of medical imaging applications such as e.g. segmentation. However, finer-scale details like tumors may be lost during translation, which is a critical limitation in oncological imaging. In this paper, we propose to address the problem of vanishing tumors for cross-modal segmentation. First, we propose a new method to insert realistic tumors in 3-D images using a deep generative model trained on a single 2-D image. Second, we leverage the proposed model using a new unpaired-then-paired two-stage I2I architecture to better penalize the suppression of tumors in cross-modal segmentation. In our experiments, we validate our model on the ongoing MICCAI crossMoDa tumor segmentation challenge, where we demonstrate superior performance over CycleGAN-based models.
Fichier principal
Vignette du fichier
MIC_sinGAN_article.pdf (282.39 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03666614 , version 1 (12-05-2022)

Identifiants

  • HAL Id : hal-03666614 , version 1

Citer

Guillaume Sallé, Pierre-Henri Conze, Nicolas Boussion, Julien Bert, Dimitris Visvikis, et al.. Synthetic tumor insertion using one-shot generative learning for cross-modal image segmentation. IEEE Nuclear science symposium and medical imaging conference 2021, Oct 2021, Virtual, Japan. ⟨hal-03666614⟩
58 Consultations
60 Téléchargements

Partager

Gmail Mastodon Facebook X LinkedIn More