Generalized Wasserstein Dice Loss, Test-Time Augmentation, and Transformers for the BraTS 2021 Challenge

TY - CHAP

T1 - Generalized Wasserstein Dice Loss, Test-Time Augmentation, and Transformers for the BraTS 2021 Challenge

AU - Fidon, Lucas

AU - Shit, Suprosanna

AU - Ezhov, Ivan

AU - Paetzold, Johannes C.

AU - Ourselin, Sébastien

AU - Vercauteren, Tom

N1 - Funding Information: Acknowledgments. This project has received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Sk lodowska-Curie grant agreement TRABIT No 765148; Wellcome [203148/Z/16/Z; WT101957], EPSRC [NS/A000049/1; NS/A000027/1]. Tom Vercauteren is supported by a Medtronic/RAEng Research Chair [RCSRF1819\7\34]. Data used in this publication were obtained as part of the RSNA-ASNR-MICCAI Brain Tumor Segmentation (BraTS) Challenge project through Synapse ID (syn25829067). Funding Information: This project has received funding from the European Union’s Horizon 2020 research and innovation program under the Marie Sk̷lodowska-Curie grant agreement TRABIT No 765148; Wellcome [203148/Z/16/Z; WT101957], EPSRC [NS/A000049/1; NS/A000027/1]. Tom Vercauteren is supported by a Medtronic/RAEng Research Chair [RCSRF1819\7\34]. Data used in this publication were obtained as part of the RSNA-ASNR-MICCAI Brain Tumor Segmentation (BraTS) Challenge project through Synapse ID (syn25829067). Publisher Copyright: © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

PY - 2022

Y1 - 2022

N2 - Brain tumor segmentation from multiple Magnetic Resonance Imaging (MRI) modalities is a challenging task in medical image computation. The main challenges lie in the generalizability to a variety of scanners and imaging protocols. In this paper, we explore strategies to increase model robustness without increasing inference time. Towards this aim, we explore finding a robust ensemble from models trained using different losses, optimizers, and train-validation data split. Importantly, we explore the inclusion of a transformer in the bottleneck of the U-Net architecture. While we find transformer in the bottleneck performs slightly worse than the baseline U-Net in average, the generalized Wasserstein Dice loss consistently produces superior results. Further, we adopt an efficient test time augmentation strategy for faster and robust inference. Our final ensemble of seven 3D U-Nets with test-time augmentation produces an average dice score of 89.4% and an average Hausdorff 95% distance of 10.0 mm when evaluated on the BraTS 2021 testing dataset. Our code and trained models are publicly available at https://github.com/LucasFidon/TRABIT_BraTS2021.

AB - Brain tumor segmentation from multiple Magnetic Resonance Imaging (MRI) modalities is a challenging task in medical image computation. The main challenges lie in the generalizability to a variety of scanners and imaging protocols. In this paper, we explore strategies to increase model robustness without increasing inference time. Towards this aim, we explore finding a robust ensemble from models trained using different losses, optimizers, and train-validation data split. Importantly, we explore the inclusion of a transformer in the bottleneck of the U-Net architecture. While we find transformer in the bottleneck performs slightly worse than the baseline U-Net in average, the generalized Wasserstein Dice loss consistently produces superior results. Further, we adopt an efficient test time augmentation strategy for faster and robust inference. Our final ensemble of seven 3D U-Nets with test-time augmentation produces an average dice score of 89.4% and an average Hausdorff 95% distance of 10.0 mm when evaluated on the BraTS 2021 testing dataset. Our code and trained models are publicly available at https://github.com/LucasFidon/TRABIT_BraTS2021.

KW - Brain tumor

KW - BraTS 2021

KW - Deep learning

KW - Segmentation

KW - Test-time augmentation

KW - Transformers

UR - http://www.scopus.com/inward/record.url?scp=85135193713&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-09002-8_17

DO - 10.1007/978-3-031-09002-8_17

M3 - Conference paper

AN - SCOPUS:85135193713

SN - 9783031090011

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 187

EP - 196

BT - Brainlesion

A2 - Crimi, Alessandro

A2 - Bakas, Spyridon

PB - Springer Science and Business Media Deutschland GmbH

T2 - 7th International Brain Lesion Workshop, BrainLes 2021, held in conjunction with the Medical Image Computing and Computer Assisted Intervention, MICCAI 2021

Y2 - 27 September 2021 through 27 September 2021

ER -