Hierarchical Brain Parcellation with Uncertainty

TY - CHAP

T1 - Hierarchical Brain Parcellation with Uncertainty

AU - Graham, Mark S.

AU - Sudre, Carole H.

AU - Varsavsky, Thomas

AU - Tudosiu, Petru Daniel

AU - Nachev, Parashkev

AU - Ourselin, Sebastien

AU - Cardoso, M. Jorge

PY - 2020

Y1 - 2020

N2 - Many atlases used for brain parcellation are hierarchically organised, progressively dividing the brain into smaller sub-regions. However, state-of-the-art parcellation methods tend to ignore this structure and treat labels as if they are ‘flat’. We introduce a hierarchically-aware brain parcellation method that works by predicting the decisions at each branch in the label tree. We further show how this method can be used to model uncertainty separately for every branch in this label tree. Our method exceeds the performance of flat uncertainty methods, whilst also providing decomposed uncertainty estimates that enable us to obtain self-consistent parcellations and uncertainty maps at any level of the label hierarchy. We demonstrate a simple way these decision-specific uncertainty maps may be used to provided uncertainty-thresholded tissue maps at any level of the label tree.

AB - Many atlases used for brain parcellation are hierarchically organised, progressively dividing the brain into smaller sub-regions. However, state-of-the-art parcellation methods tend to ignore this structure and treat labels as if they are ‘flat’. We introduce a hierarchically-aware brain parcellation method that works by predicting the decisions at each branch in the label tree. We further show how this method can be used to model uncertainty separately for every branch in this label tree. Our method exceeds the performance of flat uncertainty methods, whilst also providing decomposed uncertainty estimates that enable us to obtain self-consistent parcellations and uncertainty maps at any level of the label hierarchy. We demonstrate a simple way these decision-specific uncertainty maps may be used to provided uncertainty-thresholded tissue maps at any level of the label tree.

UR - http://www.scopus.com/inward/record.url?scp=85093075395&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-60365-6_3

DO - 10.1007/978-3-030-60365-6_3

M3 - Conference paper

AN - SCOPUS:85093075395

SN - 9783030603649

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

SP - 23

EP - 31

BT - Uncertainty for Safe Utilization of Machine Learning in Medical Imaging, and Graphs in Biomedical Image Analysis - 2nd International Workshop, UNSURE 2020, and 3rd International Workshop, GRAIL 2020, Held in Conjunction with MICCAI 2020, Proceedings

A2 - Sudre, Carole H.

A2 - Fehri, Hamid

A2 - Arbel, Tal

A2 - Baumgartner, Christian F.

A2 - Dalca, Adrian

A2 - Tanno, Ryutaro

A2 - Van Leemput, Koen

A2 - Wells, William M.

A2 - Sotiras, Aristeidis

A2 - Papiez, Bartlomiej

A2 - Ferrante, Enzo

A2 - Parisot, Sarah

PB - Springer Science and Business Media Deutschland GmbH

T2 - 2nd International Workshop on Uncertainty for Safe Utilization of Machine Learning in Medical Imaging, UNSURE 2020, and the 3rd International Workshop on Graphs in Biomedical Image Analysis, GRAIL 2020, held in conjunction with the 23rd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2020

Y2 - 8 October 2020 through 8 October 2020

ER -