Impact of imperfection in medical imaging data on deep learning-based segmentation performance: An experimental study using synthesized data.

Journal: Medical physics
Published Date: Apr 29, 2023

Abstract

BACKGROUND: Clinical data used to train deep learning models are often not clean data. They can contain imperfections in both the imaging data and the corresponding segmentations.

Authors

  • Ayetullah Mehdi Güneş
    Department of Radiation Oncology, Amsterdam UMC, Amsterdam, The Netherlands.
  • Ward van Rooij
    Amsterdam UMC, Vrije Universiteit Amsterdam, Department of Radiation Oncology, Cancer Center Amsterdam, Amsterdam, the Netherlands. Electronic address: w.vanrooij@vumc.nl.
  • Sadaf Gulshad
    Faculty of Science, Universiteit van Amsterdam, Amsterdam, The Netherlands.
  • Ben Slotman
    Department of Radiation Oncology, Amsterdam UMC, Amsterdam, The Netherlands.
  • Max Dahele
    Department of Radiotherapy, VU University Medical Center, Amsterdam, The Netherlands.
  • Wilko Verbakel
    Department of Radiation Oncology, Amsterdam UMC, Amsterdam, The Netherlands.

Keywords

External Resources

Popular Topics
Recent Journals