City Research Online

Abstract

Federated Learning is becoming ubiquitous in settings where privacy and data ownership make sharing raw data infeasible. Medical imaging presents a prominent such scenario. Despite fervent interest in Federated Learning from the Medical Imaging community, there is a general lack of standardised test-beds, datasets, and challenges that can fast-track progress in the domain. The Federated Tumour Segmentation Challenge attempts to fill that gap for the task of brain tumour segmentation. For this iteration of FeTS, we present two additional dataset splits for prototyping and test how the FedMGDA+ algorithm performs on the problem. Code for this report is provided at https://github.com/siomvas/FeTS_2022.

Publication Type:	Conference or Workshop Item (Paper)
Additional Information:	This version of the contribution has been accepted for publication, after peer review (when applicable) but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: http://dx.doi.org/10.1007/978-3-031-44153-0_15. Use of this Accepted Version is subject to the publisher’s Accepted Manuscript terms of use https://www.springernature.com/gp/open-research/policies/accepted-manuscript-terms.
Publisher Keywords:	Federated Learning, Tumour Segmentation, Medical Imaging
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science R Medicine > RC Internal medicine
Departments:	School of Science & Technology School of Science & Technology > Computer Science
SWORD Depositor:	Symplectic Administrator

Export

Downloads