Logo: to the web site of Uppsala University

uu.sePublications from Uppsala University
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Equivariant Neural Networks for Biomedical Image Analysis
Uppsala University, Disciplinary Domain of Science and Technology, Mathematics and Computer Science, Department of Information Technology, Computerized Image Analysis and Human-Computer Interaction.ORCID iD: 0000-0001-7924-6211
2024 (English)Doctoral thesis, comprehensive summary (Other academic)
Description
Abstract [en]

While artificial intelligence and deep learning have revolutionized many fields in the last decade, one of the key drivers has been access to data. This is especially true in biomedical image analysis where expert annotated data is hard to come by. The combination of Convolutional Neural Networks (CNNs) with data augmentation has proven successful in increasing the amount of training data at the cost of overfitting. In this thesis, equivariant neural networks have been used to extend the equivariant properties of CNNs to more transformations than translations. The networks have been trained and evaluated on biomedical image datasets, including bright-field microscopy images of cytological samples indicating oral cancer, and transmission electron microscopy images of virus samples. By designing the networks to be equivariant to e.g. rotations, it is shown that the need for data augmentation is reduced, that less overfitting occurs, and that convergence during training is faster. Furthermore, equivariant neural networks are more data efficient than CNNs, as demonstrated by scaling laws. These benefits are not present in all problem settings and which benefits will occur is somewhat unpredictable. We have identified that the results to some extent depend on architectures, hyperparameters and datasets. Further research may broaden the performed studies to explain how the results occur with new theory.

Place, publisher, year, edition, pages
Uppsala: Acta Universitatis Upsaliensis, 2024. , p. 82
Series
Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology, ISSN 1651-6214 ; 2352
National Category
Medical Image Processing
Research subject
Computerized Image Processing
Identifiers
URN: urn:nbn:se:uu:diva-519611ISBN: 978-91-513-2004-5 (print)OAI: oai:DiVA.org:uu-519611DiVA, id: diva2:1825662
Public defence
2024-03-01, Ångströmlaboratoriet, 101121, Sonja Lyttkens, Lägerhyddsvägen 1, Uppsala, 13:15 (English)
Opponent
Supervisors
Funder
Wallenberg AI, Autonomous Systems and Software Program (WASP)Available from: 2024-02-01 Created: 2024-01-09 Last updated: 2024-02-01
List of papers
1. Replacing data augmentation with rotation-equivariant CNNs in image-based classification of oral cancer
Open this publication in new window or tab >>Replacing data augmentation with rotation-equivariant CNNs in image-based classification of oral cancer
2021 (English)Conference paper, Published paper (Refereed)
National Category
Medical Image Processing
Identifiers
urn:nbn:se:uu:diva-460520 (URN)
Conference
25th Iberoamerican Congress on Pattern Recognition, Porto, Portugal, 10 - 13 May 2021
Funder
Wallenberg AI, Autonomous Systems and Software Program (WASP)
Available from: 2021-12-07 Created: 2021-12-07 Last updated: 2024-01-09Bibliographically approved
2. Rotation-Equivariant Semantic Instance Segmentation on Biomedical Images
Open this publication in new window or tab >>Rotation-Equivariant Semantic Instance Segmentation on Biomedical Images
2022 (English)In: MEDICAL IMAGE UNDERSTANDING AND ANALYSIS, MIUA 2022 / [ed] Yang, G Aviles-Rivero, A Roberts, M Schonlieb, CB, SPRINGER INTERNATIONAL PUBLISHING AG Springer Nature, 2022, Vol. 13413, p. 283-297Conference paper, Published paper (Refereed)
Abstract [en]

Advances in image segmentation techniques, brought by convolutional neural network (CNN) architectures like U-Net, show promise for tasks such as automated cancer screening. Recently, these methods have been extended to detect different instances of the same class, which could be used to, for example, characterize individual cells in whole-slide images. Still, the amount of data needed and the number of parameters in the network are substantial. To alleviate these problems, we modify a method of semantic instance segmentation to also enforce equivariance to the p4 symmetry group of 90-degree rotations and translations. We perform four experiments on a synthetic dataset of scattered sticks and a subset of the Kaggle 2018 Data Science Bowl, the BBBC038 dataset, consisting of segmented nuclei images. Results indicate that the rotation-equivariant architecture yields similar accuracy as a baseline architecture. Furthermore, we observe that the rotation-equivariant architecture converges faster than the baseline. This is a promising step towards reducing the training time during development of methods based on deep learning.

Place, publisher, year, edition, pages
Springer NatureSPRINGER INTERNATIONAL PUBLISHING AG, 2022
Series
Lecture Notes in Computer Science, ISSN 0302-9743, E-ISSN 1611-3349
Keywords
Deep learning, Training, Convergence
National Category
Medical Image Processing
Identifiers
urn:nbn:se:uu:diva-489382 (URN)10.1007/978-3-031-12053-4_22 (DOI)000883331000022 ()978-3-031-12053-4 (ISBN)978-3-031-12052-7 (ISBN)
Conference
26th Annual Conference on Medical Image Understanding and Analysis (MIUA), JUL 27-29, 2022, Univ Cambridge, Cambridge, ENGLAND
Funder
Wallenberg AI, Autonomous Systems and Software Program (WASP)
Available from: 2022-12-28 Created: 2022-12-28 Last updated: 2024-01-15Bibliographically approved
3. Classification of Viruses in Transmission Electron Microscopy Images using Equivariant Neural Networks
Open this publication in new window or tab >>Classification of Viruses in Transmission Electron Microscopy Images using Equivariant Neural Networks
(English)Manuscript (preprint) (Other academic)
National Category
Medical Image Processing
Research subject
Computerized Image Processing
Identifiers
urn:nbn:se:uu:diva-519609 (URN)
Funder
Wallenberg AI, Autonomous Systems and Software Program (WASP)
Available from: 2024-01-08 Created: 2024-01-08 Last updated: 2024-01-09
4. Equivariant Neural Networks for Biomedical Images Improves Data Efficiency
Open this publication in new window or tab >>Equivariant Neural Networks for Biomedical Images Improves Data Efficiency
(English)Manuscript (preprint) (Other academic)
National Category
Medical Image Processing
Research subject
Computerized Image Processing
Identifiers
urn:nbn:se:uu:diva-519610 (URN)
Funder
Wallenberg AI, Autonomous Systems and Software Program (WASP)
Available from: 2024-01-08 Created: 2024-01-08 Last updated: 2024-01-09

Open Access in DiVA

UUThesis_K-Bengtsson-Bernander-2024(1455 kB)230 downloads
File information
File name FULLTEXT01.pdfFile size 1455 kBChecksum SHA-512
7b881b3c265bfb845790b439ffb4273cd2f5e729a907201249c62b892ffebb3b57301f1058721a9f4f880f5fda79ecac3071ee944df4934206273c3441727a7c
Type fulltextMimetype application/pdf

Authority records

Bengtsson Bernander, Karl

Search in DiVA

By author/editor
Bengtsson Bernander, Karl
By organisation
Computerized Image Analysis and Human-Computer Interaction
Medical Image Processing

Search outside of DiVA

GoogleGoogle Scholar
Total: 230 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

isbn
urn-nbn

Altmetric score

isbn
urn-nbn
Total: 1272 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf