Particle Tracking with PolyParticleTracker

Particle tracking software based on the "Polynomial Fit Gaussian Weight" method
by Salman S Rogers, Biological Physics Group, University of Manchester (2007)

Description

PolyParticleTracker is a software package which accurately tracks particles in microscope images, using the "Polynomial Fit Gaussian Weight" (PFGW) method of Rogers et al. (2007). PolyParticleTracker is particularly suitable for the tracking of low-contrast particles against a complicated background, e.g. endogeneous particles in cell images. Input images for PolyParticleTracker may be either uncompressed AVI movie files (single files or series of movie segments) or individual numbered frame images (e.g. TIF, PNG, BMP). The package is released free for non-commercial use; it consists of a set of Matlab scripts and must be run within the Matlab programming environment, version 7.0 or higher. PolyParticleTracker contains a graphical user interface for ease of use (screenshot below).

Download

The package can be downloaded here as a zip file, consisting of a set of Matlab scripts, figures and HTML documentation. Last updated 21 May 2008.


Download Here!

Installation instructions:

  1. Unzip polyparticletracker.zip to produce all files in a single directory "polyparticletracker".
  2. Move this directory to somewhere convenient for Matlab.
  3. Start Matlab and add the polyparticletracker directory to your Matlab path (File menu - Set Path). All scripts and help files should now be accessible at the Matlab command prompt. (Usage instructions below.)

How PolyParticleTracker works

PolyParticleTracker implements the "Polynomial Fit Gaussian Weight" method: after smoothing each image with a noise-reduction filter, the intensity at each particle is fitted with a quartic polynomial. This fit is weighted by circular Gaussian function of decay length equal to the estimated particle radius. This radius is, in turn, calculated from the polynomial fit. Since the particle position is based on a polynomial fit and the background is excluded with a Gaussian weighting function, it is inherently insensitive to nearby variations in the background (e.g. intensity surface below), and allows some variation in the shape of a particle (e.g. for tracking ellipsoids or other non-spherical particles).

Examples

Three examples with captions are shown below: left to right. Detailed examples can be found in Rogers et al. (2007).
1. Bright field image of L929 fibroblasts (with 10 micron scale bar). Tracks of endogeneous particles are superposed. Many tracks show clear active motion (examples are highlighted green). Many tracks show stepwise motion along their length - e.g. 36 nm steps corresponding to myosin motion along actin filaments (below). 2. Fluorescent image of 0.5-micron polymer beads in HeLa cells. Particle tracks are superposed (white). 3. Streaming pseudopod of Amoeba proteus (with 10 micron scale bar) (a). Flow is marked by many endogeneous particles of different sizes and shapes. Particle tracks (b) are obtained with PolyParticleTracker.

Instructions

Reference and further information

Salman Rogers welcomes comments about the software or method, and any bugs that might be found. If PolyParticleTracker has been useful to you, we would be grateful if you would cite our paper describing its method in any work that has made use of it:

Rogers S.S., Waigh T.A., Zhao X. and Lu J.R., Physical Biology (2007) "Precise Particle Tracking Against a Complicated Background: Polynomial Fitting with Gaussian Weight". Publication: http://www.iop.org/EJ/abstract/1478-3975/4/3/007 Preprint: http://arxiv.org/abs/0707.3602

Link

See also Eric Weeks's website Particle tracking using IDL, with more tracking methods and further information.