Details
| Original language | English |
|---|---|
| Pages (from-to) | 10022-10030 |
| Number of pages | 9 |
| Journal | Environmental Science & Technology |
| Volume | 54 |
| Issue number | 16 |
| Early online date | 14 Jul 2020 |
| Publication status | Published - 18 Aug 2020 |
Abstract
While the use of deep learning is a valuable technology for automatic detection systems for medical data and images, the biofouling community is still lacking an analytical tool for the detection and counting of diatoms on samples after short-term field exposure. In this work, a fully convolutional neural network was implemented as a fast and simple approach to detect diatoms on two-channel (fluorescence and phase-contrast) microscopy images by predicting bounding boxes. The developed approach performs well with only a small number of trainable parameters and a F1 score of 0.82. Counting diatoms was evaluated on a data set of 600 microscopy images of three different surface chemistries (hydrophilic and hydrophobic) and is very similar to counting by humans while demanding only a fraction of the analysis time.
ASJC Scopus subject areas
- Chemistry(all)
- General Chemistry
- Environmental Science(all)
- Environmental Chemistry
Cite this
- Standard
- Harvard
- Apa
- Vancouver
- BibTeX
- RIS
In: Environmental Science & Technology, Vol. 54, No. 16, 18.08.2020, p. 10022-10030.
Research output: Contribution to journal › Article › Research › peer review
}
TY - JOUR
T1 - Fully Convolutional Neural Network for Detection and Counting of Diatoms on Coatings after Short-Term Field Exposure
AU - Krause, Lutz M. K.
AU - Koc, Julian
AU - Rosenhahn, Bodo
AU - Rosenhahn, Axel
N1 - Funding information: The authors are grateful to the Deutsche Forschungsgemeinschaft (DFG, RO 2524/4-1 and RO 2524/4-2) and the Office of Naval Research (ONR, Grants N00014-16-12979 and N00014-20-12244) for providing financial support.
PY - 2020/8/18
Y1 - 2020/8/18
N2 - While the use of deep learning is a valuable technology for automatic detection systems for medical data and images, the biofouling community is still lacking an analytical tool for the detection and counting of diatoms on samples after short-term field exposure. In this work, a fully convolutional neural network was implemented as a fast and simple approach to detect diatoms on two-channel (fluorescence and phase-contrast) microscopy images by predicting bounding boxes. The developed approach performs well with only a small number of trainable parameters and a F1 score of 0.82. Counting diatoms was evaluated on a data set of 600 microscopy images of three different surface chemistries (hydrophilic and hydrophobic) and is very similar to counting by humans while demanding only a fraction of the analysis time.
AB - While the use of deep learning is a valuable technology for automatic detection systems for medical data and images, the biofouling community is still lacking an analytical tool for the detection and counting of diatoms on samples after short-term field exposure. In this work, a fully convolutional neural network was implemented as a fast and simple approach to detect diatoms on two-channel (fluorescence and phase-contrast) microscopy images by predicting bounding boxes. The developed approach performs well with only a small number of trainable parameters and a F1 score of 0.82. Counting diatoms was evaluated on a data set of 600 microscopy images of three different surface chemistries (hydrophilic and hydrophobic) and is very similar to counting by humans while demanding only a fraction of the analysis time.
UR - http://www.scopus.com/inward/record.url?scp=85089711253&partnerID=8YFLogxK
U2 - 10.1021/acs.est.0c01982
DO - 10.1021/acs.est.0c01982
M3 - Article
C2 - 32663392
AN - SCOPUS:85089711253
VL - 54
SP - 10022
EP - 10030
JO - Environmental Science & Technology
JF - Environmental Science & Technology
SN - 0013-936X
IS - 16
ER -