Generalized Wasserstein Dice Score, Distributionally Robust Deep Learning, and Ranger for Brain Tumor Segmentation: BraTS 2020 Challenge

Lucas Fidon*, Sébastien Ourselin, Tom Vercauteren

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference paperpeer-review

32 Citations (Scopus)

Abstract

Training a deep neural network is an optimization problem with four main ingredients: the design of the deep neural network, the per-sample loss function, the population loss function, and the optimizer. However, methods developed to compete in recent BraTS challenges tend to focus only on the design of deep neural network architectures, while paying less attention to the three other aspects. In this paper, we experimented with adopting the opposite approach. We stuck to a generic and state-of-the-art 3D U-Net architecture and experimented with a non-standard per-sample loss function, the generalized Wasserstein Dice loss, a non-standard population loss function, corresponding to distributionally robust optimization, and a non-standard optimizer, Ranger. Those variations were selected specifically for the problem of multi-class brain tumor segmentation. The generalized Wasserstein Dice loss is a per-sample loss function that allows taking advantage of the hierarchical structure of the tumor regions labeled in BraTS. Distributionally robust optimization is a generalization of empirical risk minimization that accounts for the presence of underrepresented subdomains in the training dataset. Ranger is a generalization of the widely used Adam optimizer that is more stable with small batch size and noisy labels. We found that each of those variations of the optimization of deep neural networks for brain tumor segmentation leads to improvements in terms of Dice scores and Hausdorff distances. With an ensemble of three deep neural networks trained with various optimization procedures, we achieved promising results on the validation dataset and the testing dataset of the BraTS 2020 challenge. Our ensemble ranked fourth out of 78 for the segmentation task of the BraTS 2020 challenge with mean Dice scores of 88.9, 84.1, and 81.4, and mean Hausdorff distances at 95 % of 6.4, 19.4, and 15.8 for the whole tumor, the tumor core, and the enhancing tumor.

Original languageEnglish
Title of host publicationBrainlesion
Subtitle of host publicationGlioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 6th International Workshop, BrainLes 2020, Held in Conjunction with MICCAI 2020, Revised Selected Papers
EditorsAlessandro Crimi, Spyridon Bakas
PublisherSpringer Science and Business Media Deutschland GmbH
Pages200-214
Number of pages15
ISBN (Print)9783030720865
DOIs
Publication statusPublished - 2021
Event6th International MICCAI Brainlesion Workshop, BrainLes 2020 Held in Conjunction with 23rd Medical Image Computing for Computer Assisted Intervention Conference, MICCAI 2020 - Virtual, Online
Duration: 4 Oct 20204 Oct 2020

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12659 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference6th International MICCAI Brainlesion Workshop, BrainLes 2020 Held in Conjunction with 23rd Medical Image Computing for Computer Assisted Intervention Conference, MICCAI 2020
CityVirtual, Online
Period4/10/20204/10/2020

Keywords

  • Brain tumor
  • BraTS challenge
  • Convolutional neural network
  • Dice score
  • Distributionally robust optimization
  • Segmentation

Fingerprint

Dive into the research topics of 'Generalized Wasserstein Dice Score, Distributionally Robust Deep Learning, and Ranger for Brain Tumor Segmentation: BraTS 2020 Challenge'. Together they form a unique fingerprint.

Cite this