Tackling the class imbalance problem of deep learning-based head and neck organ segmentation - Takeaways - MDSpire

Tackling the class imbalance problem of deep learning-based head and neck organ segmentation

  • By

  • Elias Tappeiner

  • Martin Welk

  • Rainer Schubert

  • May 16, 2022

  • 0 min

Share

  • 1

    Head and neck tumors account for 3% of newly diagnosed cancers, necessitating precise organ segmentation for effective radiotherapy.

  • 2

    Deep learning approaches are increasingly used for automated segmentation of head and neck organs, addressing the time-consuming manual process.

  • 3

    Class imbalance in segmentation arises from the large size differences between organs and the ratio of background to foreground voxels.

  • 4

    Optimizing training patch sizes and adapting the Dice loss function can enhance segmentation performance in deep learning models.

  • 5

    Recent advancements include hybrid networks and combined loss functions that effectively address class imbalance in head and neck organ segmentation.

Original Source(s)

Tackling the class imbalance problem of deep learning-based head and neck organ segmentation

  • by Elias Tappeiner, Martin Welk, Rainer Schubert

  • May 16, 2022

Related Content

Assessment of the mechanical performance of a novel CT-linac treatment couch

FDA Grants Accelerated Approval for OTOF Gene Therapy

A one-time intracochlear gene therapy improved hearing outcomes in patients with OTOF-related sensorineural hearing loss in the phase 1/2 CHORD trial.

Efficacy Comparison of Neoadjuvant Short-Course and Long-Course Radiotherapy Regimens With or Without Immunotherapy for Locally Advanced pMMR Rectal Cancer: A Systematic Review and Network Meta-Analysis