An iterative algorithm to improve colloidal particle locating

TL;DR

This paper introduces an iterative algorithm that enhances the accuracy and reliability of locating colloidal particles in dense 3D images, addressing issues like missed detections and double counting, and is adaptable to varying image brightness.

Contribution

The proposed algorithm improves particle locating accuracy in dense colloids, reduces parameter sensitivity, and is compatible with existing software, advancing imaging analysis in soft matter physics.

Findings

Reduces missed particle detections

Minimizes double counting errors

Effective with spatially-varying brightness images

Abstract

Confocal microscopy of colloids combined with digital image processing has become a powerful tool in soft matter physics and materials science. Together, these techniques enable locating and tracking of more than half a million individual colloidal particles at once. However, despite improvements in locating algorithms that improve position accuracy, it remains challenging to locate all particles in a densely-packed, three dimensional colloid without erroneously identifying the same particle more than once. We present a simple, iterative algorithm that mitigates both the "missed particle" and "double counting" problems while simultaneously reducing sensitivity to the specific choice of input parameters. It is also useful for analyzing images with spatially-varying brightness in which a single set of input parameters is not appropriate for all particles. The algorithm is easy to implement and compatible with existing particle locating software.