Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Novel MRI-US fusion with advanced annotation in focal cryoablation for prostate cancer: medium-term outcomes and learning curve insights.

BJU international·2025
Same author

First-line atezolizumab/bevacizumab or durvalumab/tremelimumab in advanced hepatocellular carcinoma: a real world, multicenter retrospective study.

The oncologist·2025
Same author

In-Bore MRI-Guided Ureteral Stent Placement During Prostate Cancer Cryoablation-A Case Series.

Diagnostics (Basel, Switzerland)·2025
Same author

Photon counting detector CTA for prostate artery embolization pre-procedure planning and intra-procedural guidance.

CVIR endovascular·2025
Same author

Research-based clinical deployment of artificial intelligence algorithm for prostate MRI.

Abdominal radiology (New York)·2025
Same author

The influence of the pharmaceutical industry on the development of gonadotrophins and ovarian stimulation protocols in assisted reproductive technologies.

Frontiers in endocrinology·2025

Related Experiment Video

Updated: Jul 24, 2025

Label-Free Identification of Lymphocyte Subtypes Using Three-Dimensional Quantitative Phase Imaging and Machine Learning
08:58

Label-Free Identification of Lymphocyte Subtypes Using Three-Dimensional Quantitative Phase Imaging and Machine Learning

Published on: November 19, 2018

12.6K

Novel machine learning models for flow imaging microscopy sub-visible particle classification in protein

Robert Bassett1, Dharmini Mehta1, Scott Thompson1

  • 1CSL Innovation, 655 Elizabeth St, Melbourne, 3000, VIC, Australia.

International Journal of Pharmaceutics
|July 4, 2023
PubMed
Summary

This study enhances particle analysis in drug products using machine learning. Techniques like data augmentation and transfer learning improve classification accuracy, even with limited data, ensuring drug safety.

Keywords:
Flow imaging microscopyNeural networksSub-visible particles

More Related Videos

Detecting and Characterizing Protein Self-Assembly In Vivo by Flow Cytometry
05:58

Detecting and Characterizing Protein Self-Assembly In Vivo by Flow Cytometry

Published on: July 17, 2019

11.1K
Confocal Imaging of Confined Quiescent and Flowing Colloid-polymer Mixtures
10:56

Confocal Imaging of Confined Quiescent and Flowing Colloid-polymer Mixtures

Published on: May 20, 2014

12.2K

Related Experiment Videos

Last Updated: Jul 24, 2025

Label-Free Identification of Lymphocyte Subtypes Using Three-Dimensional Quantitative Phase Imaging and Machine Learning
08:58

Label-Free Identification of Lymphocyte Subtypes Using Three-Dimensional Quantitative Phase Imaging and Machine Learning

Published on: November 19, 2018

12.6K
Detecting and Characterizing Protein Self-Assembly In Vivo by Flow Cytometry
05:58

Detecting and Characterizing Protein Self-Assembly In Vivo by Flow Cytometry

Published on: July 17, 2019

11.1K
Confocal Imaging of Confined Quiescent and Flowing Colloid-polymer Mixtures
10:56

Confocal Imaging of Confined Quiescent and Flowing Colloid-polymer Mixtures

Published on: May 20, 2014

12.2K

Area of Science:

  • Pharmaceutical Sciences
  • Analytical Chemistry
  • Computational Biology

Background:

  • Accurate particulate content analysis is crucial for drug product safety and efficacy.
  • Traditional particle counting methods lack classification capabilities, failing to distinguish harmful aggregated proteins or silicone oil droplets.
  • Flow imaging microscopy combined with machine learning (ML) offers simultaneous particle classification and counting.

Purpose of the Study:

  • To explore techniques for enhancing prediction accuracy in particle classification using ML models.
  • To address challenges associated with limited labeled datasets in training ML models for particle analysis.
  • To improve the identification and differentiation of various particle types in pharmaceutical formulations.

Main Methods:

  • Utilized flow imaging microscopy for particle visualization and data acquisition.
  • Applied machine learning (ML) models, including convolutional neural networks (CNNs), for particle classification.
  • Investigated strategies to improve model performance with limited labeled data, including data augmentation and transfer learning.
  • Explored novel models integrating both imaging and tabular data for enhanced classification.

Main Results:

  • Demonstrated that combining data augmentation and transfer learning significantly improves ML model prediction accuracy for particle classification.
  • Showcased the effectiveness of novel models that integrate imaging and tabular data for superior particle identification.
  • Achieved high prediction accuracy even when training datasets were limited, a common challenge in this field.

Conclusions:

  • The integration of advanced ML techniques, including data augmentation and transfer learning, is vital for accurate particle classification in drug products.
  • Novel approaches combining imaging and tabular data offer a promising direction for robust particle analysis.
  • These advancements contribute to improved patient safety by enabling better characterization of particulate matter in pharmaceuticals.