Computer vision for high-throughput analysis of pickering emulsions.

Area of Science:

• Materials Science
• Chemical Engineering
• Image Analysis

Background:

• Pickering emulsions, stabilized by solid particles, are vital in food, cosmetics, and pharmaceuticals.
• Quantitative droplet size analysis is challenging due to non-uniform distribution and polydispersity in microscopy images.

Purpose of the Study:

• To investigate the circle-Hough transform (CHT) for quantitative droplet size assessment in Pickering emulsions.
• To develop and validate an open-source software tool, Hough-Scan, for this analysis.

Main Methods:

• Application of the circle-Hough transform (CHT) for detecting circular features in microscopy images.
• Development of the open-source software 'Hough-Scan' with a graphical interface and tiling algorithm.
• Comparison of CHT performance against manual identification and other computer vision methods.

Main Results:

• The CHT, implemented via Hough-Scan, demonstrated superior precision, recall, and accuracy in identifying and sizing Pickering emulsion droplets.
• Hough-Scan effectively handles variations in spatial distribution, size, and visual artifacts in microscopy images.
• The software's tiling algorithm enhances computational efficiency for large datasets.

Conclusions:

• The CHT, facilitated by Hough-Scan, significantly improves the ease and accuracy of Pickering emulsion droplet size analysis.
• This approach is applicable to diverse Pickering emulsion systems and common microscopy image challenges.
• Hough-Scan provides a user-friendly and efficient solution for quantitative emulsion analysis.