General Defocusing Particle Tracking: Fundamentals and uncertainty assessment

Rune Barnkob, Massimiliano Rossi

Published 2019-12-03Version 1

General Defocusing Particle Tracking (GDPT) refers to a class of three-dimensional particle tracking methods that use a single-camera view and determine the particle depth positions from the defocusing patterns of the corresponding particle images. Its distinctive feature is to access particles' depth coordinates through a direct comparison of the particle images with a set of reference particle images at known depth positions. Many implementations of GDPT are possible, especially with respect to how to compare target and calibration images, how to deal with overlapping particle images, and how to optimize the computational time. The emergence of new and more sophisticated GDPT approaches requires the definition of the method fundamentals as well as a standardized framework for the objective assessment of the method performance and applicability. To meet this need, we identify and describe the fundamental concepts and parameters defining GDPT. We define guidelines for a standardized assessment of the efficiency and uncertainty of GDPT implementations. In particular, we show that a complete GDPT assessment must state the obtained particle detection rate, the depth coordinate uncertainty, and the measurement depth. In addition, we provide datasets for the evaluation of a GDPT implementation's dependency on image noise, particle image overlapping, and light intensity variations. We used the presented datasets and guidelines to assess the performance of DefocusTracker, a freely-accessible GDPT implementation based on the normalized cross-correlation. The assessment of DefocusTracker illustrates fundamental concepts of GDPT analysis and paves the road as a first benchmark for further development of GDPT.