Automated tracking of colloidal clusters with sub-pixel accuracy and precision

TL;DR

This paper introduces an advanced particle tracking method capable of accurately tracking overlapping colloidal features with sub-pixel precision, enabling detailed analysis of colloidal clusters and their dynamics.

Contribution

The authors develop a novel tracking algorithm that incorporates feature history and non-linear least-squares fitting, improving accuracy and precision over existing methods.

Findings

Tracking accuracy below 0.2% of feature radius

Precision of 0.1 to 0.01 pixels in typical images

Successful extraction of 3D diffusion tensor from colloidal dimers

Abstract

Quantitative tracking of features from video images is a basic technique employed in many areas of science. Here, we present a method for the tracking of features that partially overlap, in order to be able to track so-called colloidal molecules. Our approach implements two improvements into existing particle tracking algorithms. Firstly, we use the history of previously identified feature locations to successfully find their positions in consecutive frames. Secondly, we present a framework for non-linear least-squares fitting to summed radial model functions and analyze the accuracy (bias) and precision (random error) of the method on artificial data. We find that our tracking algorithm correctly identifies overlapping features with an accuracy below 0.2% of the feature radius and a precision of 0.1 to 0.01 pixels for a typical image of a colloidal cluster. Finally, we use our method to extract the three-dimensional diffusion tensor from the Brownian motion of colloidal dimers.