uu.seUppsala University Publications
Change search
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
Deep Neural Networks and Image Analysis for Quantitative Microscopy
Uppsala University, Disciplinary Domain of Science and Technology, Mathematics and Computer Science, Department of Information Technology, Division of Visual Information and Interaction. Uppsala University, Disciplinary Domain of Science and Technology, Mathematics and Computer Science, Department of Information Technology, Computerized Image Analysis and Human-Computer Interaction. SciLifeLab. (Visual information and interaction)
2017 (English)Doctoral thesis, comprehensive summary (Other academic)
Description
Abstract [en]

Understanding biology paves the way for discovering drugs targeting deadly diseases like cancer, and microscopy imaging is one of the most informative ways to study biology. However, analysis of large numbers of samples is often required to draw statistically verifiable conclusions. Automated approaches for analysis of microscopy image data makes it possible to handle large data sets, and at the same time reduce the risk of bias. Quantitative microscopy refers to computational methods for extracting measurements from microscopy images, enabling detection and comparison of subtle changes in morphology or behavior induced by varying experimental conditions. This thesis covers computational methods for segmentation and classification of biological samples imaged by microscopy.

Recent increase in computational power has enabled the development of deep neural networks (DNNs) that perform well in solving real world problems. This thesis compares classical image analysis algorithms for segmentation of bacteria cells and introduces a novel method that combines classical image analysis and DNNs for improved cell segmentation and detection of rare phenotypes. This thesis also demonstrates a novel DNN for segmentation of clusters of cells (spheroid), with varying sizes, shapes and textures imaged by phase contrast microscopy. DNNs typically require large amounts of training data. This problem is addressed by proposing an automated approach for creating ground truths by utilizing multiple imaging modalities and classical image analysis. The resulting DNNs are applied to segment unstained cells from bright field microscopy images. In DNNs, it is often difficult to understand what image features have the largest influence on the final classification results. This is addressed in an experiment where DNNs are applied to classify zebrafish embryos based on phenotypic changes induced by drug treatment. The response of the trained DNN is tested by ablation studies, which revealed that the networks do not necessarily learn the features most obvious at visual examination. Finally, DNNs are explored for classification of cervical and oral cell samples collected for cancer screening. Initial results show that the DNNs can respond to very subtle malignancy associated changes. All the presented methods are developed using open-source tools and validated on real microscopy images.

Place, publisher, year, edition, pages
Uppsala: Acta Universitatis Upsaliensis, 2017. , 85 p.
Series
Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology, ISSN 1651-6214 ; 1566
Keyword [en]
Deep neural networks, convolutional neural networks, image analysis, quantitative microscopy, bright-field microscopy
National Category
Other Electrical Engineering, Electronic Engineering, Information Engineering Signal Processing
Research subject
Computerized Image Processing
Identifiers
URN: urn:nbn:se:uu:diva-329834ISBN: 978-91-513-0080-1 (print)OAI: oai:DiVA.org:uu-329834DiVA: diva2:1143597
Public defence
2017-11-10, 2446, ITC, Lägerhyddsvägen 2, Hus 2, Uppsala, 10:15 (English)
Opponent
Supervisors
Funder
Swedish Research Council, 2012-4968EU, European Research Council, 682810eSSENCE - An eScience Collaboration
Available from: 2017-10-17 Created: 2017-09-21 Last updated: 2017-10-17
List of papers
1. Segmentation and track-analysis in time-lapse imaging of bacteria
Open this publication in new window or tab >>Segmentation and track-analysis in time-lapse imaging of bacteria
Show others...
2016 (English)In: IEEE Journal on Selected Topics in Signal Processing, ISSN 1932-4553, E-ISSN 1941-0484, Vol. 10, no 1, 174-184 p.Article in journal (Refereed) Published
Abstract [en]

In this paper, we have developed tools to analyze prokaryotic cells growing in monolayers in a microfluidic device. Individual bacterial cells are identified using a novel curvature based approach and tracked over time for several generations. The resulting tracks are thereafter assessed and filtered based on track quality for subsequent analysis of bacterial growth rates. The proposed method performs comparable to the state-of-the-art methods for segmenting phase contrast and fluorescent images, and we show a 10-fold increase in analysis speed.

Keyword
E. coli; microscopy; segmentation; time-lapse; tracking
National Category
Bioinformatics (Computational Biology)
Research subject
Computerized Image Processing
Identifiers
urn:nbn:se:uu:diva-265457 (URN)10.1109/JSTSP.2015.2491304 (DOI)000369495900015 ()
Projects
eSSENCE
Funder
Swedish Research Council, 2012-4968EU, European Research Council, 616047eSSENCE - An eScience Collaboration
Available from: 2016-01-21 Created: 2015-10-29 Last updated: 2017-11-17
2. Fast Adaptive Local Thresholding Based on Ellipse fit
Open this publication in new window or tab >>Fast Adaptive Local Thresholding Based on Ellipse fit
2016 (English)Conference paper, Published paper (Refereed)
Abstract [en]

In this paper we propose an adaptive thresholding method where each object is thresholded optimizing its shape. The method is based on a component tree representation, which can be computed in quasi-linear time. We test and evaluate the method on images of bacteria from three different live-cell analysis experiments and show that the proposed method produces segmentation results comparable to state-of-the-art but at least an order of magnitude faster. The method can be extended to compute any feature measurements that can be calculated in a cumulative way, and holds great potential for applications where a priori information on expected object size and shape is available.

Series
Scripta minora Bibliothecae regiae Universitatis Upsaliensis, ISSN 0282-3152
Series
IEEE International Symposium on Biomedical Imaging, ISSN 1945-7928
National Category
Medical Image Processing
Research subject
Computerized Image Processing
Identifiers
urn:nbn:se:uu:diva-275182 (URN)000386377400049 ()9781479923496 (ISBN)9781479923502 (ISBN)
Conference
International Symposium on Biomedical Imaging (ISBI'16), Prague, Czech Republic, April 13-16, 2016
Funder
Science for Life Laboratory - a national resource center for high-throughput molecular bioscienceSwedish Research Council, 2012-4968
Available from: 2016-02-01 Created: 2016-02-01 Last updated: 2017-09-21Bibliographically approved
3. Feature augmented deep neural networks for segmentation of cells
Open this publication in new window or tab >>Feature augmented deep neural networks for segmentation of cells
2016 (English)In: Computer Vision – ECCV 2016 Workshops: Part I, Springer, 2016, 231-243 p.Conference paper, Published paper (Refereed)
Place, publisher, year, edition, pages
Springer, 2016
Series
Lecture Notes in Computer Science, 9913
National Category
Medical Image Processing
Research subject
Computerized Image Processing
Identifiers
urn:nbn:se:uu:diva-306914 (URN)10.1007/978-3-319-46604-0_17 (DOI)978-3-319-46603-3 (ISBN)
Conference
IEEE International Conference on Computer Vision (ICCV)
Funder
Swedish Research Council, 2012-4968
Available from: 2016-09-18 Created: 2016-11-05 Last updated: 2017-11-17
4. Spheroid segmentation using multiscale deep adversarial networks
Open this publication in new window or tab >>Spheroid segmentation using multiscale deep adversarial networks
2017 (English)In: IEEE International Conference on Computer Vision, 2017Conference paper, Poster (with or without abstract) (Refereed)
National Category
Medical Image Processing
Research subject
Computerized Image Processing
Identifiers
urn:nbn:se:uu:diva-329355 (URN)
Conference
ICCV Workshop on Bioimage Computing, Venice, Italy, October 23, 2017.
Funder
Swedish Research Council, 2012-4968EU, European Research Council, 682810
Available from: 2017-09-13 Created: 2017-09-13 Last updated: 2017-11-17
5. Automated training of deep convolutional neural networks for cell segmentation
Open this publication in new window or tab >>Automated training of deep convolutional neural networks for cell segmentation
2017 (English)In: Scientific Reports, ISSN 2045-2322, E-ISSN 2045-2322, Vol. 7, 7860Article in journal (Refereed) Published
National Category
Medical Image Processing
Research subject
Computerized Image Processing
Identifiers
urn:nbn:se:uu:diva-329301 (URN)10.1038/s41598-017-07599-6 (DOI)28798336 (PubMedID)
Funder
Swedish Research Council, 2012-4968EU, European Research Council, 682810
Available from: 2017-08-10 Created: 2017-09-12 Last updated: 2017-11-17
6. Deep Fish: Deep Learning-Based Classification of Zebrafish Deformation for High-Throughput Screening
Open this publication in new window or tab >>Deep Fish: Deep Learning-Based Classification of Zebrafish Deformation for High-Throughput Screening
2017 (English)In: Journal of Biomolecular Screening, ISSN 1087-0571, E-ISSN 1552-454X, Vol. 22, no 1, 102-107 p.Article in journal (Refereed) Published
Abstract [en]

Zebrafish (Danio rerio) is an important vertebrate model organism in biomedical research, especially suitable for morphological screening due to its transparent body during early development. Deep learning has emerged as a dominant paradigm for data analysis and found a number of applications in computer vision and image analysis. Here we demonstrate the potential of a deep learning approach for accurate high-throughput classification of whole-body zebrafish deformations in multifish microwell plates. Deep learning uses the raw image data as an input, without the need of expert knowledge for feature design or optimization of the segmentation parameters. We trained the deep learning classifier on as few as 84 images (before data augmentation) and achieved a classification accuracy of 92.8% on an unseen test data set that is comparable to the previous state of the art (95%) based on user-specified segmentation and deformation metrics. Ablation studies by digitally removing whole fish or parts of the fish from the images revealed that the classifier learned discriminative features from the image foreground, and we observed that the deformations of the head region, rather than the visually apparent bent tail, were more important for good classification performance.

National Category
Signal Processing
Research subject
Computerized Image Processing
Identifiers
urn:nbn:se:uu:diva-309535 (URN)10.1177/1087057116667894 (DOI)000394206000012 ()27613194 (PubMedID)
Funder
Swedish Research Council, 2012-4968eSSENCE - An eScience Collaboration
Available from: 2016-12-05 Created: 2016-12-05 Last updated: 2017-11-17
7. Deep convolutional neural networks for detecting cellular changes due to malignancy
Open this publication in new window or tab >>Deep convolutional neural networks for detecting cellular changes due to malignancy
Show others...
2017 (English)In: IEEE International Conference on Computer Vision, 2017Conference paper, Poster (with or without abstract) (Refereed)
National Category
Medical Image Processing
Research subject
Computerized Image Processing
Identifiers
urn:nbn:se:uu:diva-329356 (URN)
Conference
ICCV workshop on Bioimage Computing, Venice, Italy, October 23, 2017.
Funder
EU, European Research Council, 682810Swedish Research Council, 2012-4968
Available from: 2017-09-13 Created: 2017-09-13 Last updated: 2017-11-17

Open Access in DiVA

fulltext(1584 kB)216 downloads
File information
File name FULLTEXT01.pdfFile size 1584 kBChecksum SHA-512
2f4b80bc334ddcb1503f72cd6b1ec83af77f5194dbfdb3ba4dbdc2dcd7ab3c8a2e3639a3ad4609410bb62581060e5d10458d320fe70ba7b56207024a5ccfce15
Type fulltextMimetype application/pdf

By organisation
Division of Visual Information and InteractionComputerized Image Analysis and Human-Computer Interaction
Other Electrical Engineering, Electronic Engineering, Information EngineeringSignal Processing

Search outside of DiVA

GoogleGoogle Scholar
Total: 216 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: 1269 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