Umeå University's logo

umu.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • 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
A hybrid CNN-Random Forest algorithm for bacterial spore segmentation and classification in TEM images
Umeå University, Faculty of Science and Technology, Department of Physics.
Umeå University, Faculty of Science and Technology, Department of Physics.ORCID iD: 0000-0002-0168-0197
Umeå University, Faculty of Science and Technology, Department of Physics.ORCID iD: 0000-0002-0496-6692
Umeå University, Faculty of Science and Technology, Department of Physics. Umeå University, Faculty of Medicine, Umeå Centre for Microbial Research (UCMR). (The Biophysics and Biophotonics group)ORCID iD: 0000-0002-9835-3263
2023 (English)In: Scientific Reports, E-ISSN 2045-2322, Vol. 13, no 1, article id 18758Article in journal (Refereed) Published
Abstract [en]

We present a new approach to segment and classify bacterial spore layers from Transmission Electron Microscopy (TEM) images using a hybrid Convolutional Neural Network (CNN) and Random Forest (RF) classifier algorithm. This approach utilizes deep learning, with the CNN extracting features from images, and the RF classifier using those features for classification. The proposed model achieved 73% accuracy, 64% precision, 46% sensitivity, and 47% F1-score with test data. Compared to other classifiers such as AdaBoost, XGBoost, and SVM, our proposed model demonstrates greater robustness and higher generalization ability for non-linear segmentation. Our model is also able to identify spores with a damaged core as verified using TEMs of chemically exposed spores. Therefore, the proposed method will be valuable for identifying and characterizing spore features in TEM images, reducing labor-intensive work as well as human bias.

Place, publisher, year, edition, pages
Springer Nature, 2023. Vol. 13, no 1, article id 18758
National Category
Other Physics Topics Other Computer and Information Science
Identifiers
URN: urn:nbn:se:umu:diva-216165DOI: 10.1038/s41598-023-44212-5ISI: 001123935800008PubMedID: 37907463Scopus ID: 2-s2.0-85175591485OAI: oai:DiVA.org:umu-216165DiVA, id: diva2:1809565
Funder
Swedish Research Council, 2019-04016The Kempe Foundations, JCK-2129.3Available from: 2023-11-04 Created: 2023-11-04 Last updated: 2025-09-30Bibliographically approved
In thesis
1. Spotlight the killer: detecting harmful chemical and biological agents using optical spectroscopy
Open this publication in new window or tab >>Spotlight the killer: detecting harmful chemical and biological agents using optical spectroscopy
2025 (English)Doctoral thesis, comprehensive summary (Other academic)
Alternative title[sv]
Lyset på mördaren : detektion av skadliga kemiska och biologiska ämnen med hjälp av optisk spektroskopi
Abstract [en]

Harmful chemical and biological agents are a significant threat to health and prosperity worldwide. Recent years have seen an increase in wars and conflicts around the globe, raising concerns about the potential deployment of chemical and biological warfare agents. On a less speculative level, harmful chemicals such as narcotic substances cause immense humanitarian and economic damage through overdoses and associated healthcare costs, while microbes such as pathogenic bacteria and parasites cause hospital-acquired infections and food spoilage at a cost of approximately 1 trillion euros every year. To combat the threat of these harmful agents, we must thus develop rapid and effective detection and diagnostic methods for harmful agents, allowing us to effectively deploy specific treatments and preventative measures.

Classically, while there exist numerous methods for the detection of both harmful chemical and biological agents, they often come with limitations that inhibit their effectiveness. These inhibitions often take the form of bulky equipment that is difficult to apply in the field or time-consuming preparation and measurement processes.

In this thesis we will explore an alternative category of assays for detecting and characterizing harmful materials – optical spectroscopy. Optical spectroscopy is a category of material characterization methods that use light to probe a material. While probing the material, we receive a signal characteristic of the molecules, chemical, and biological structure of our material. These optical spectroscopic methods, such as Raman spectroscopy and fluorescence spectroscopy, can be used to characterize a material within the span of minutes or even seconds, making them ideal for detection applications. Furthermore, they can often be made portable or even handheld, making them a great tool for initial field indication of harmful materials, ahead of thorough lab analysis.

I sincerely hope the studies presented herein can serve as a stepping stone to future technologies and detection assays, capable of saving both money and lives. 

Place, publisher, year, edition, pages
Umeå: Umeå University, 2025. p. 72
Keywords
Sensing, Raman spectroscopy, SERS, Fluorescence spectroscopy, CWA, nerve agents, bacterial spores, Cryptosporidium
National Category
Atom and Molecular Physics and Optics
Identifiers
urn:nbn:se:umu:diva-244830 (URN)978-91-8070-780-0 (ISBN)978-91-8070-779-4 (ISBN)
Public defence
2025-10-24, Aula Anatomica, Biologihuset, 907 36, Umeå, Umeå, 13:00 (English)
Opponent
Supervisors
Note

This work was done in collaboration with, and with support from, the Swedish Defece Research Agency (FOI).

Available from: 2025-10-03 Created: 2025-09-30 Last updated: 2025-10-22Bibliographically approved

Open Access in DiVA

fulltext(4043 kB)176 downloads
File information
File name FULLTEXT01.pdfFile size 4043 kBChecksum SHA-512
a212685efac4cfc04036123a8f25122057a85a7b2e1dcda79719f1463c458ae34f04fb2a6b1752abef403eb88bbbe7602e095229389ecf776063d10b9920f101
Type fulltextMimetype application/pdf

Other links

Publisher's full textPubMedScopus

Authority records

Qamar, SaqibÖberg, RasmusMalyshev, DmitryAndersson, Magnus

Search in DiVA

By author/editor
Qamar, SaqibÖberg, RasmusMalyshev, DmitryAndersson, Magnus
By organisation
Department of PhysicsUmeå Centre for Microbial Research (UCMR)
In the same journal
Scientific Reports
Other Physics TopicsOther Computer and Information Science

Search outside of DiVA

GoogleGoogle Scholar
Total: 177 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

doi
pubmed
urn-nbn

Altmetric score

doi
pubmed
urn-nbn
Total: 565 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • 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