Efficient Bayesian Uncertainty Estimation for nnU-Net

Efficient Bayesian Uncertainty Estimation for nnU-Net

Zhao, Y. (TU Delft ImPhys/Medical Imaging)
Yang, C. (TU Delft ImPhys/Medical Imaging)
Schweidtmann, A.M. (TU Delft ChemE/Product and Process Engineering)
Tao, Q. (TU Delft ImPhys/Medical Imaging)

Wang, Linwei (editor)
Dou, Qi (editor)
Fletcher, P. Thomas (editor)
Speidel, Stefanie (editor)
Li, Shuo (editor)

2022

The self-configuring nnU-Net has achieved leading performance in a large range of medical image segmentation challenges. It is widely considered as the model of choice and a strong baseline for medical image segmentation. However, despite its extraordinary performance, nnU-Net does not supply a measure of uncertainty to indicate its possible failure. This can be problematic for large-scale image segmentation applications, where data are heterogeneous and nnU-Net may fail without notice. In this work, we introduce a novel method to estimate nnU-Net uncertainty for medical image segmentation. We propose a highly effective scheme for posterior sampling of weight space for Bayesian uncertainty estimation. Different from previous baseline methods such as Monte Carlo Dropout and mean-field Bayesian Neural Networks, our proposed method does not require a variational architecture and keeps the original nnU-Net architecture intact, thereby preserving its excellent performance and ease of use. Additionally, we boost the segmentation performance over the original nnU-Net via marginalizing multi-modal posterior models. We applied our method on the public ACDC and M &M datasets of cardiac MRI and demonstrated improved uncertainty estimation over a range of baseline methods. The proposed method further strengthens nnU-Net for medical image segmentation in terms of both segmentation accuracy and quality control.

nnU-Net
Stochastic gradient descent
Uncertainty estimation
Variational inference

http://resolver.tudelft.nl/uuid:938b8ec5-13ab-4f90-801b-529cc8608d66

https://doi.org/10.1007/978-3-031-16452-1_51

Springer

2023-07-01

978-3-031-16451-4

Medical Image Computing and Computer Assisted Intervention – MICCAI 2022 - 25th International Conference, Proceedings

25th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2022, 2022-09-18 → 2022-09-22, Singapore, Singapore

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 0302-9743, 13438 LNCS

Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

Institutional Repository

conference paper

© 2022 Y. Zhao, C. Yang, A.M. Schweidtmann, Q. Tao