Guiding Unsupervised MRI-to-CT synthesis using Content and style Representation by an Enhanced Perceptual synthesis (CREPs) loss - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

Guiding Unsupervised MRI-to-CT synthesis using Content and style Representation by an Enhanced Perceptual synthesis (CREPs) loss

Résumé

The goal of this research was to propose an unsupervised learning technique for producing synthetic CT (sCT) images from MRI data. For model training, a dataset consisting of 180 pairs of brain CT and MR scans, as well as 180 pairs of pelvis scans was used. The devised methodology incorporates a 3D conditional Generative Adversarial Network (cGAN) training in an unsupervised way. To tackle challenges associated with unsupervised learning convergence, a novel ConvNext-based perceptual loss (CREPs loss) was developed to guide in the 3D cGAN-based MR-to-CT generation process.
Fichier principal
Vignette du fichier
Method_Description_MRI.pdf (201.23 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04272482 , version 1 (06-11-2023)

Identifiants

  • HAL Id : hal-04272482 , version 1

Citer

Cédric Hémon, Valentin Boussot, Blanche Texier, Jean-Louis Dillenseger, Jean-Claude Nunes. Guiding Unsupervised MRI-to-CT synthesis using Content and style Representation by an Enhanced Perceptual synthesis (CREPs) loss. SynthRAD2023 Challenge, MICCAI 2023,, Oct 2023, Vancouver (BC), Canada. ⟨hal-04272482⟩
22 Consultations
9 Téléchargements

Partager

Gmail Facebook X LinkedIn More