Domain generalization in deep learning for contrast-enhanced imaging.

Journal: Computers in biology and medicine
Published Date:

Abstract

BACKGROUND: The domain generalization problem has been widely investigated in deep learning for non-contrast imaging over the last years, but it received limited attention for contrast-enhanced imaging. However, there are marked differences in contrast imaging protocols across clinical centers, in particular in the time between contrast injection and image acquisition, while access to multi-center contrast-enhanced image data is limited compared to available datasets for non-contrast imaging. This calls for new tools for generalizing single-domain, single-center deep learning models across new unseen domains and clinical centers in contrast-enhanced imaging.

Authors

  • Carla Sendra-Balcells
    Dept. de Matemàtiques i Informàtica, Universitat de Barcelona, Spain. Electronic address: carla.sendra@ub.edu.
  • Víctor M Campello
    Dept. de Matemàtiques i Informàtica, Universitat de Barcelona, Spain.
  • Carlos Martín-Isla
    Dept. de Matemàtiques i Informàtica, Universitat de Barcelona, Spain.
  • David Vilades
    Institut de Recerca Sant Pau (IR SANT PAU), Sant Quintí 77-79, 08041, Barcelona, Spain.
  • Martín L Descalzo
    Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona, Spain.
  • Andrea Guala
    Cardiovascular Imaging Unit, Hospital Universitari Vall d'Hebron, Barcelona, Spain.
  • José F Rodríguez-Palomares
    Cardiovascular Imaging Unit, Hospital Universitari Vall d'Hebron, Barcelona, Spain.
  • Karim Lekadir
    Information and Communication Technologies Department, Universitat Pompeu Fabra, Barcelona, Spain.