Please use this identifier to cite or link to this item:
http://hdl.handle.net/1942/34686
Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.author | ROUSSEAU, Axel-Jan | - |
dc.contributor.author | BECKER, Thijs | - |
dc.contributor.author | Bertels, Jeroen | - |
dc.contributor.author | Blaschko, Matthew B. | - |
dc.contributor.author | VALKENBORG, Dirk | - |
dc.date.accessioned | 2021-08-19T15:32:49Z | - |
dc.date.available | 2021-08-19T15:32:49Z | - |
dc.date.issued | 2021 | - |
dc.date.submitted | 2021-08-12T10:53:27Z | - |
dc.identifier.citation | 2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI), p. 1052 -1056 | - |
dc.identifier.isbn | 978-1-6654-1246-9 | - |
dc.identifier.issn | 1945-7928 | - |
dc.identifier.uri | http://hdl.handle.net/1942/34686 | - |
dc.description.abstract | Neural networks for automated image segmentation are typically trained to achieve maximum accuracy, while less attention has been given to the calibration of their confidence scores. However, well-calibrated confidence scores provide valuable information towards the user. We investigate several post hoc calibration methods that are straightforward to implement, some of which are novel. They are compared to Monte Carlo (MC) dropout and are applied to neural networks trained with cross-entropy (CE) and soft Dice (SD) losses on BraTS 2018 and ISLES 2018. Surprisingly, models trained on SD loss are not necessarily less calibrated than those trained on CE loss. In all cases, at least one post hoc method improves the calibration. There is limited consistency across the results, so we can't conclude on one method being superior. In all cases, post hoc calibration is competitive with MC dropout. Although average calibration improves compared to the base model, subject-level variance of the calibration remains similar. | - |
dc.language.iso | en | - |
dc.publisher | IEEE | - |
dc.relation.ispartofseries | IEEE International Symposium on Biomedical Imaging | - |
dc.subject.other | uncertainty estimation | - |
dc.subject.other | confidence calibration | - |
dc.subject.other | deep learning | - |
dc.subject.other | Medical image segmentation | - |
dc.title | Post Training Uncertainty Calibration Of Deep Networks For Medical Image Segmentation | - |
dc.type | Proceedings Paper | - |
local.bibliographicCitation.conferencedate | 13-16 April 2021 | - |
local.bibliographicCitation.conferencename | 2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI) | - |
local.bibliographicCitation.conferenceplace | Nice, France | - |
dc.identifier.epage | 1056 | - |
dc.identifier.spage | 1052 | - |
local.bibliographicCitation.jcat | C1 | - |
local.publisher.place | 345 E 47TH ST, NEW YORK, NY 10017 USA | - |
local.type.refereed | Refereed | - |
local.type.specified | Proceedings Paper | - |
local.type.programme | VSC | - |
dc.identifier.doi | 10.1109/ISBI48211.2021.9434131 | - |
dc.identifier.isi | 000786144100220 | - |
local.provider.type | CrossRef | - |
local.bibliographicCitation.btitle | 2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI) | - |
local.uhasselt.uhpub | yes | - |
local.uhasselt.international | no | - |
item.contributor | ROUSSEAU, Axel-Jan | - |
item.contributor | BECKER, Thijs | - |
item.contributor | Bertels, Jeroen | - |
item.contributor | Blaschko, Matthew B. | - |
item.contributor | VALKENBORG, Dirk | - |
item.fullcitation | ROUSSEAU, Axel-Jan; BECKER, Thijs; Bertels, Jeroen; Blaschko, Matthew B. & VALKENBORG, Dirk (2021) Post Training Uncertainty Calibration Of Deep Networks For Medical Image Segmentation. In: 2021 IEEE 18th International Symposium on Biomedical Imaging (ISBI), p. 1052 -1056. | - |
item.accessRights | Open Access | - |
item.fulltext | With Fulltext | - |
item.validation | ecoom 2023 | - |
Appears in Collections: | Research publications |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
POST_TRAINING_UNCERTAINTY_CALIBRATION_OF_DEEP_NETWORKS_FOR_MEDICAL_IMAGE_SEGMENTATION.pdf Restricted Access | Published version | 1.65 MB | Adobe PDF | View/Open Request a copy |
4545_arXiv.pdf | Non Peer-reviewed author version | 1.13 MB | Adobe PDF | View/Open |
WEB OF SCIENCETM
Citations
10
checked on May 2, 2024
Page view(s)
68
checked on Sep 7, 2022
Download(s)
14
checked on Sep 7, 2022
Google ScholarTM
Check
Altmetric
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.