ConText-GAN: using contextual texture information for realistic and controllable medical image synthesis* - Aix-Marseille Université Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

ConText-GAN: using contextual texture information for realistic and controllable medical image synthesis*

Résumé

This study proposes an enhancement to the ConText-GAN, an image synthesis model using a controllable texture input. The improvement consists in using a texture feature fusion module to reduce the complexity of the model, and enable the use of the OASIS architecture for image generation. Clinical relevance— The ConText-GAN can be used to generate images of fake patients, which are useful in the medical field due to the scarcity of data. An example is given of the generation of images showing pathological muscle tissue in the context of neuromuscular diseases.
Fichier principal
Vignette du fichier
Hostin_BHI2023.pdf (6.53 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04350740 , version 1 (13-03-2024)

Identifiants

Citer

Marc-Adrien Hostin, Shahram Attarian, David Bendahan, Marc-Emmanuel Bellemare. ConText-GAN: using contextual texture information for realistic and controllable medical image synthesis*. 2023 IEEE EMBS International Conference on Biomedical and Health Informatics (BHI), IEEE EMBS, Oct 2023, Pittsburgh, United States. ⟨10.1109/BHI58575.2023.10313436⟩. ⟨hal-04350740⟩
42 Consultations
4 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More