Related Experiment Video
Updated: Sep 18, 2025
09:56
In Vivo Multimodal Imaging and Analysis of Mouse Laser-Induced Choroidal Neovascularization Model
Published on: January 21, 2018
9.4K
Harnessing Real-Time UV Imaging and Convolutional Neural Networks (CNNs): Unlocking New Opportunities for Empirical
Maciej Stróżyk1, Adam Pacławski1, Aleksander Mendyk1
1Department of Pharmaceutical Technology and Biopharmaceutics, Jagiellonian University Medical College, 30-688 Kraków, Poland.
Pharmaceutics
|June 27, 2025
Summary
This study explores using UV imaging and AI models to predict drug performance. Convolutional neural networks (CNNs) show promise in analyzing dissolution data directly from images for better drug development.
Area of Science:
- Pharmaceutical Sciences
- Computational Chemistry
- Biomedical Engineering
Background:
- Investigating real-time UV imaging for drug dissolution analysis.
- Utilizing convolutional neural networks (CNNs) for multidimensional modeling.
- Bridging the gap between in vitro drug dissolution and in vivo performance.
Purpose of the Study:
- To develop predictive models for drug performance using advanced imaging and AI.
- To evaluate the efficacy of traditional machine learning versus deep learning approaches.
- To explore the potential of CNNs in capturing complex dissolution patterns directly from imaging data.
Main Methods:
- Employing the SDi2 apparatus for multidimensional dissolution data capture.
- Studying Glucophage tablets (immediate and extended-release) in various media and wavelengths.
- Applying traditional machine learning (Scikit-learn, Tensorflow, AutoML) and CNNs on imaging data.
Main Results:
- Acquired comprehensive dissolution data at 255 nm and 520 nm.
- Compared performance of models using extracted numerical data versus raw images.
- Demonstrated CNNs' ability to predict in vivo metformin plasma concentrations directly from SDi2 images.
Conclusions:
- Real-time UV imaging combined with AI offers a powerful approach to drug performance modeling.
- CNNs can potentially uncover hidden information in dissolution images missed by traditional analysis.
- This dual approach enhances understanding of drug dissolution and its in vivo implications.
