Academic Bibliography

 Search 200 years of publications by Ghent University researchers.
Advanced search
2 files | 9.24 MB
Add to list
  1. Search results
  2. Evaluation of activated sludge settli...

Evaluation of activated sludge settling characteristics from microscopy images with deep convolutional neural networks and transfer learning

Sina Borzooei (UGent) , Leonardo Scabini, Gisele Miranda, Saba Daneshgar (UGent) , Lukas Deblieck, Odemir Bruno, Piet De Langhe, Bernard De Baets (UGent) , Ingmar Nopens (UGent) and Elena Torfs (UGent)
(2024) JOURNAL OF WATER PROCESS ENGINEERING. 64.
Author
Organization
Project
Abstract
Timely assessment and prediction of changes in microbial compositions leading to activated sludge settling problems, such as filamentous bulking (FB), can reduce water resource recovery facilities (WRRFs) upsets, operational challenges, and negative environmental impacts. This study presents a computer vision approach to assess activated sludge-settling characteristics based on Microscopy Images (MIs). We utilize MIs to train deep convolutional neural networks (CNN) using transfer learning to investigate the morphological properties of flocs and filaments. The methodology was tested on the offline MI dataset collected over two years at a full-scale industrial WRRF in Belgium. Various CNN architectures were tested, including Inception v3, ResNet18, ResNet152, ConvNeXt-nano, and ConvNeXt-S. The sludge volume index (SVI) was used as the final prediction variable, but the method can be easily adjusted to predict any other settling metric of choice. The bestperforming CNN, ConvNeXt-nano, could predict SVI values with MAE (37.51 +/- 4.02), MTD (11.65 +/- 1.94), MAPE (0.18 +/- 0.02), and R 2 (0.75 +/- 0.05). The model was tested in real-life FB events, where it identified early indicators of bulking by predictive surges in SVI values. We used an explainable AI technique, Eigen-CAM, to discover key morphological indicators of sludge bulking transitions. The findings highlight the SVI multimodality issue, where SVI readings as a unidimensional metric could not capture delicate shifts from good to poor sludge settling, while the model detected these subtle changes. The key morphological attributes of threshold conditions leading to FB were identified, which can provide actionable insight for preemptive WRRF management.
Keywords
Wastewater treatment plant, Filamentous bulking, Convolutional neural networks, Transfer learning, Microscopy images, Eigen-CAM, WASTE-WATER TREATMENT, FILAMENTOUS BACTERIA, BULKING, CLASSIFICATION, MORPHOLOGY, SEPARATION, FLOCS

Downloads

  • Download
    (...).pdf
    • full text (Published version)
    • |
    • UGent only
    • |
    • PDF
    • |
    • 7.06 MB
  • Download
    Computer Vision Manuscript Sina Borzooei Revised Not Marked.pdf
    • full text (Accepted manuscript)
    • |
    • open access
    • |
    • PDF
    • |
    • 2.18 MB

Citation

Please use this url to cite or link to this publication:

MLA
Borzooei, Sina, et al. “Evaluation of Activated Sludge Settling Characteristics from Microscopy Images with Deep Convolutional Neural Networks and Transfer Learning.” JOURNAL OF WATER PROCESS ENGINEERING, vol. 64, 2024, doi:10.1016/j.jwpe.2024.105692.
APA
Borzooei, S., Scabini, L., Miranda, G., Daneshgar, S., Deblieck, L., Bruno, O., … Torfs, E. (2024). Evaluation of activated sludge settling characteristics from microscopy images with deep convolutional neural networks and transfer learning. JOURNAL OF WATER PROCESS ENGINEERING, 64. https://doi.org/10.1016/j.jwpe.2024.105692
Chicago author-date
Borzooei, Sina, Leonardo Scabini, Gisele Miranda, Saba Daneshgar, Lukas Deblieck, Odemir Bruno, Piet De Langhe, Bernard De Baets, Ingmar Nopens, and Elena Torfs. 2024. “Evaluation of Activated Sludge Settling Characteristics from Microscopy Images with Deep Convolutional Neural Networks and Transfer Learning.” JOURNAL OF WATER PROCESS ENGINEERING 64. https://doi.org/10.1016/j.jwpe.2024.105692.
Chicago author-date (all authors)
Borzooei, Sina, Leonardo Scabini, Gisele Miranda, Saba Daneshgar, Lukas Deblieck, Odemir Bruno, Piet De Langhe, Bernard De Baets, Ingmar Nopens, and Elena Torfs. 2024. “Evaluation of Activated Sludge Settling Characteristics from Microscopy Images with Deep Convolutional Neural Networks and Transfer Learning.” JOURNAL OF WATER PROCESS ENGINEERING 64. doi:10.1016/j.jwpe.2024.105692.
Vancouver
1.
Borzooei S, Scabini L, Miranda G, Daneshgar S, Deblieck L, Bruno O, et al. Evaluation of activated sludge settling characteristics from microscopy images with deep convolutional neural networks and transfer learning. JOURNAL OF WATER PROCESS ENGINEERING. 2024;64.
IEEE
[1]
S. Borzooei et al., “Evaluation of activated sludge settling characteristics from microscopy images with deep convolutional neural networks and transfer learning,” JOURNAL OF WATER PROCESS ENGINEERING, vol. 64, 2024.
@article{01J1VTHGF5XRM5JG21EC0XD9YX,
  abstract     = {{Timely assessment and prediction of changes in microbial compositions leading to activated sludge settling problems, such as filamentous bulking (FB), can reduce water resource recovery facilities (WRRFs) upsets, operational challenges, and negative environmental impacts. This study presents a computer vision approach to assess activated sludge-settling characteristics based on Microscopy Images (MIs). We utilize MIs to train deep convolutional neural networks (CNN) using transfer learning to investigate the morphological properties of flocs and filaments. The methodology was tested on the offline MI dataset collected over two years at a full-scale industrial WRRF in Belgium. Various CNN architectures were tested, including Inception v3, ResNet18, ResNet152, ConvNeXt-nano, and ConvNeXt-S. The sludge volume index (SVI) was used as the final prediction variable, but the method can be easily adjusted to predict any other settling metric of choice. The bestperforming CNN, ConvNeXt-nano, could predict SVI values with MAE (37.51 +/- 4.02), MTD (11.65 +/- 1.94), MAPE (0.18 +/- 0.02), and R 2 (0.75 +/- 0.05). The model was tested in real-life FB events, where it identified early indicators of bulking by predictive surges in SVI values. We used an explainable AI technique, Eigen-CAM, to discover key morphological indicators of sludge bulking transitions. The findings highlight the SVI multimodality issue, where SVI readings as a unidimensional metric could not capture delicate shifts from good to poor sludge settling, while the model detected these subtle changes. The key morphological attributes of threshold conditions leading to FB were identified, which can provide actionable insight for preemptive WRRF management.}},
  articleno    = {{105692}},
  author       = {{Borzooei, Sina and Scabini, Leonardo and Miranda, Gisele and Daneshgar, Saba and Deblieck, Lukas and Bruno, Odemir and De Langhe, Piet and De Baets, Bernard and Nopens, Ingmar and Torfs, Elena}},
  issn         = {{2214-7144}},
  journal      = {{JOURNAL OF WATER PROCESS ENGINEERING}},
  keywords     = {{Wastewater treatment plant,Filamentous bulking,Convolutional neural networks,Transfer learning,Microscopy images,Eigen-CAM,WASTE-WATER TREATMENT,FILAMENTOUS BACTERIA,BULKING,CLASSIFICATION,MORPHOLOGY,SEPARATION,FLOCS}},
  language     = {{eng}},
  pages        = {{13}},
  title        = {{Evaluation of activated sludge settling characteristics from microscopy images with deep convolutional neural networks and transfer learning}},
  url          = {{http://doi.org/10.1016/j.jwpe.2024.105692}},
  volume       = {{64}},
  year         = {{2024}},
}
Altmetric
View in Altmetric
Web of Science
Times cited: