Semixup : in- and out-of-manifold regularization for deep semi-supervised knee osteoarthritis severity grading from plain radiographs |
|
Author: | Nguyen, Huy Hoang1; Saarakkala, Simo2; Blaschko, Matthew B.3; |
Organizations: |
1Research Unit of Medical Imaging, Physics and Technology, University of Oulu, Oulu, Finland 2Research Unit of Medical Imaging, Physics and Technology, University of Oulu, Oulu, Finland Oulu University Hospital, Oulu, Finland 3Center for Processing Speech and Images, KU Leuven, Leuven, Belgium
4Ailean Technologies Oy, Oulu, Finland
|
Format: | article |
Version: | published version |
Access: | open |
Online Access: | PDF Full Text (PDF, 2.1 MB) |
Persistent link: | http://urn.fi/urn:nbn:fi-fe202101192144 |
Language: | English |
Published: |
Institute of Electrical and Electronics Engineers,
2020
|
Publish Date: | 2021-01-19 |
Description: |
AbstractKnee osteoarthritis (OA) is one of the highest disability factors in the world. This musculoskeletal disorder is assessed from clinical symptoms, and typically confirmed via radiographic assessment. This visual assessment done by a radiologist requires experience, and suffers from moderate to high inter-observer variability. The recent literature has shown that deep learning methods can reliably perform the OA severity assessment according to the gold standard Kellgren-Lawrence (KL) grading system. However, these methods require large amounts of labeled data, which are costly to obtain. In this study, we propose the Semixup algorithm, a semi-supervised learning (SSL) approach to leverage unlabeled data. Semixup relies on consistency regularization using in- and out-of-manifold samples, together with interpolated consistency. On an independent test set, our method significantly outperformed other state-of-the-art SSL methods in most cases. Finally, when compared to a well-tuned fully supervised baseline that yielded a balanced accuracy (BA) of 70.9 ± 0.8% on the test set, Semixup had comparable performance - BA of 71 ± 0.8% (p = 0.368) while requiring 6 times less labeled data. These results show that our proposed SSL method allows building fully automatic OA severity assessment tools with datasets that are available outside research settings. see all
|
Series: |
IEEE transactions on medical imaging |
ISSN: | 0278-0062 |
ISSN-E: | 1558-254X |
ISSN-L: | 0278-0062 |
Volume: | 39 |
Issue: | 12 |
Pages: | 4346 - 4356 |
DOI: | 10.1109/TMI.2020.3017007 |
OADOI: | https://oadoi.org/10.1109/TMI.2020.3017007 |
Type of Publication: |
A1 Journal article – refereed |
Field of Science: |
217 Medical engineering 3126 Surgery, anesthesiology, intensive care, radiology |
Subjects: | |
Funding: |
Private sector funding for the OAI is managed by the Foundation for the National Institutes of Health. |
Copyright information: |
© 2020 The Authors. This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/. |
https://creativecommons.org/licenses/by/4.0/ |