Different binarization processes validated against manual counts of fluorescent bacterial cells
| Autor: | Gijsbert J. Jansen, Astrid H. Paulitsch-Fuchs, Gerrit G. Tamminga, Gert-Jan Euverink |
|---|---|
| Přispěvatelé: | Products and Processes for Biotechnology |
| Jazyk: | angličtina |
| Rok vydání: | 2016 |
| Předmět: |
0301 basic medicine
Microbiology (medical) DNA Bacterial 030106 microbiology Analytical chemistry Colony Count Microbial Value (computer science) Magnification SOFTWARE Biology Binarization Microbiology Grayscale Fluorescence 03 medical and health sciences Software Enumeration Image Processing Computer-Assisted PARTICLES Automated-cell-count ENUMERATION IMAGE-ANALYSIS Molecular Biology LASER-SCANNING MICROSCOPY Background subtraction Reproducibility Bacteriological Techniques Microscopy NUMBERS Bacteria business.industry Binary image Reproducibility of Results Pattern recognition LABELED BACTERIA AUTOMATIC-DETERMINATION Algorithm RNA Bacterial 030104 developmental biology Microscopy Fluorescence Artificial intelligence business Algorithms |
| Zdroj: | Journal of microbiological methods, 128, 118-124. ELSEVIER SCIENCE BV |
| ISSN: | 0167-7012 |
| DOI: | 10.1016/j.mimet.2016.07.003 |
| Popis: | State of the art software methods (such as fixed value approaches or statistical approaches) to create a binary image of fluorescent bacterial cells are not as accurate and precise as they should be for counting bacteria and measuring their area. To overcome these bottlenecks, we introduce biological significance to obtain a binary image from a greyscale microscopic image. Using our biological significance approach we are able to automatically count about the same number of cells as an individual researcher would do by manual/visual counting. Using the fixed value or statistical approach to obtain a binary image leads to about 20% less cells in automatic counting. In our procedure we included the area measurements of the bacterial cells to determine the right parameters for background subtraction and threshold values. In an iterative process the threshold and background subtraction values were incremented until the number of particles smaller than a typical bacterial cell is less than the number of bacterial cells with a certain area. This research also shows that every image has a specific threshold with respect to the optical system, magnification and staining procedure as well as the exposure time. The biological significance approach shows that automatic counting can be performed with the same accuracy, precision and reproducibility as manual counting. The same approach can be used to count bacterial cells using different optical systems (Leica, Olympus and Navitar), magnification factors (200x and 400x), staining procedures (DNA (Propidium Iodide) and RNA (FISH)) and substrates (polycarbonate filter or glass). (C) 2016 Elsevier B.V. All rights reserved. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect