Academic Bibliography

 Search 200 years of publications by Ghent University researchers.
Advanced search
2 files | 10.33 MB
Add to list
  1. Search results
  2. A deep learning approach to perform d...

A deep learning approach to perform defect classification of freeze-dried product

Quentin Hervé (UGent) , Nusret Ipek (UGent) , Jan Verwaeren (UGent) and Thomas De Beer (UGent)
(2025) INTERNATIONAL JOURNAL OF PHARMACEUTICS. 670.
Author
Organization
Project
Abstract
Cosmetic inspection of freeze-dried products is an important part of the post-manufacturing quality control process. Traditionally done by human visual inspection, this method poses typical challenges and shortcomings that can be addressed with innovative techniques. While many cosmetic defects can occur, some are considered more critical than others as they can be harmful to the patient or affect the drug's efficacy. With the rise of artificial intelligence and computer vision technology, faster and more reproducible quality control is possible, allowing real-time monitoring on a continuous manufacturing line. In this study, several continuously freezedried samples were prepared using formulations and process settings that lead deliberately to specific defects faced in freeze-drying as well as defect-free samples. Two approaches (i.e. patch-based approach and multi- label classification) capable of handling high-resolution images based on Convolutional Neural Networks were developed and compared to select the optimal one. Additional visualization techniques were used to enhance model understanding further. The best approach achieved perfect precision and recall on critical defects, with a prediction time of less than 50 ms to make a decision on the acceptance or rejection of vials generated.
Keywords
Freeze-drying, Continuous manufacturing, Quality analysis, Visual inspection, Deep learning, Computer vision, Multi-class classification, Multi-label classification, INSPECTION, STABILITY, COLLAPSE

Downloads

  • Download
    Defect Classification Journal Version.pdf
    • full text (Accepted manuscript)
    • |
    • open access
    • |
    • PDF
    • |
    • 8.14 MB
  • Download
    (...).pdf
    • full text (Published version)
    • |
    • UGent only
    • |
    • PDF
    • |
    • 2.20 MB

Citation

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

MLA
Hervé, Quentin, et al. “A Deep Learning Approach to Perform Defect Classification of Freeze-Dried Product.” INTERNATIONAL JOURNAL OF PHARMACEUTICS, vol. 670, 2025, doi:10.1016/j.ijpharm.2024.125127.
APA
Hervé, Q., Ipek, N., Verwaeren, J., & De Beer, T. (2025). A deep learning approach to perform defect classification of freeze-dried product. INTERNATIONAL JOURNAL OF PHARMACEUTICS, 670. https://doi.org/10.1016/j.ijpharm.2024.125127
Chicago author-date
Hervé, Quentin, Nusret Ipek, Jan Verwaeren, and Thomas De Beer. 2025. “A Deep Learning Approach to Perform Defect Classification of Freeze-Dried Product.” INTERNATIONAL JOURNAL OF PHARMACEUTICS 670. https://doi.org/10.1016/j.ijpharm.2024.125127.
Chicago author-date (all authors)
Hervé, Quentin, Nusret Ipek, Jan Verwaeren, and Thomas De Beer. 2025. “A Deep Learning Approach to Perform Defect Classification of Freeze-Dried Product.” INTERNATIONAL JOURNAL OF PHARMACEUTICS 670. doi:10.1016/j.ijpharm.2024.125127.
Vancouver
1.
Hervé Q, Ipek N, Verwaeren J, De Beer T. A deep learning approach to perform defect classification of freeze-dried product. INTERNATIONAL JOURNAL OF PHARMACEUTICS. 2025;670.
IEEE
[1]
Q. Hervé, N. Ipek, J. Verwaeren, and T. De Beer, “A deep learning approach to perform defect classification of freeze-dried product,” INTERNATIONAL JOURNAL OF PHARMACEUTICS, vol. 670, 2025.
@article{01JMA5NS5VQF9KQRRJ06AKZ2PJ,
  abstract     = {{Cosmetic inspection of freeze-dried products is an important part of the post-manufacturing quality control process. Traditionally done by human visual inspection, this method poses typical challenges and shortcomings that can be addressed with innovative techniques. While many cosmetic defects can occur, some are considered more critical than others as they can be harmful to the patient or affect the drug's efficacy. With the rise of artificial intelligence and computer vision technology, faster and more reproducible quality control is possible, allowing real-time monitoring on a continuous manufacturing line. In this study, several continuously freezedried samples were prepared using formulations and process settings that lead deliberately to specific defects faced in freeze-drying as well as defect-free samples. Two approaches (i.e. patch-based approach and multi- label classification) capable of handling high-resolution images based on Convolutional Neural Networks were developed and compared to select the optimal one. Additional visualization techniques were used to enhance model understanding further. The best approach achieved perfect precision and recall on critical defects, with a prediction time of less than 50 ms to make a decision on the acceptance or rejection of vials generated.}},
  articleno    = {{125127}},
  author       = {{Hervé, Quentin and Ipek, Nusret and Verwaeren, Jan and De Beer, Thomas}},
  issn         = {{0378-5173}},
  journal      = {{INTERNATIONAL JOURNAL OF PHARMACEUTICS}},
  keywords     = {{Freeze-drying,Continuous manufacturing,Quality analysis,Visual inspection,Deep learning,Computer vision,Multi-class classification,Multi-label classification,INSPECTION,STABILITY,COLLAPSE}},
  language     = {{eng}},
  pages        = {{11}},
  title        = {{A deep learning approach to perform defect classification of freeze-dried product}},
  url          = {{http://doi.org/10.1016/j.ijpharm.2024.125127}},
  volume       = {{670}},
  year         = {{2025}},
}
Altmetric
View in Altmetric
Web of Science
Times cited: