This is an archive of the old MediaWiki-based ImageJ wiki. The current website can be found at imagej.net.

Using TrackMate from MATLAB

This page shows how to use and control TrackMate directly from within MATLAB. This is great and made possible thanks to the great Miji tool, that make the Fiji classes visible from MATLAB. Check this page first if you have not already. Note however that as 2016, Mark Hinerm and friends built a stronger replacement from Miji, ImageJ-MATLAB. This page still clings to using Miji, but moving to ImageJ-MATLAB should be painless.

All the following examples assume you have launched MATLAB, and properly initiated Miji, using for instance

Miji(false)

or

Miji(true)

if you want the Fiji window to be visible.

Apart from that, the collections of examples that follow just look like several scripts. Which is actually the case: We use the MATLAB interpreter as a scripting interface for TrackMate, as we did in the Scripting TrackMate page. We separated this page to highlight MATLAB specificities and to exploit its capabilities better.


A simple tracking example

Here we open an image though Fiji (not though MATLAB) and track its content. The results are displayed in a Fiji window as well. MATLAB is used here only to control TrackMate and does not really plays a role in the process. This example is actually the same that in the first Scripting TrackMate example.

Note that we used the fully qualified name for TrackMate classes, e.g. fiji.plugin.trackmate.TrackMate instead of TrackMate. This is the first way to import classes in a MATLAB script. We will see another way later.

% Get currently selected image
% imp = ij.IJ.openImage('https://fiji.sc/samples/FakeTracks.tif')
imp = ij.ImagePlus('/Users/tinevez/Desktop/Data/FakeTracks.tif');
imp.show()
   
   
%----------------------------
% Create the model object now
%----------------------------
   
% Some of the parameters we configure below need to have
% a reference to the model at creation. So we create an
% empty model now.
model = fiji.plugin.trackmate.Model();
   
% Send all messages to ImageJ log window.
model.setLogger(fiji.plugin.trackmate.Logger.IJ_LOGGER)
      
%------------------------
% Prepare settings object
%------------------------
      
settings = fiji.plugin.trackmate.Settings();
settings.setFrom(imp)
      
% Configure detector - We use a java map
settings.detectorFactory = fiji.plugin.trackmate.detection.LogDetectorFactory();
map = java.util.HashMap();
map.put('DO_SUBPIXEL_LOCALIZATION', true);
map.put('RADIUS', 2.5);
map.put('TARGET_CHANNEL', 1);
map.put('THRESHOLD', 0);
map.put('DO_MEDIAN_FILTERING', false);
settings.detectorSettings = map;
   
% Configure spot filters - Classical filter on quality
filter1 = fiji.plugin.trackmate.features.FeatureFilter('QUALITY', 1.0, true);
settings.addSpotFilter(filter1)
    
% Configure tracker - We want to allow splits and fusions
settings.trackerFactory  = fiji.plugin.trackmate.tracking.sparselap.SparseLAPTrackerFactory();
settings.trackerSettings = fiji.plugin.trackmate.tracking.LAPUtils.getDefaultLAPSettingsMap(); % almost good enough
settings.trackerSettings.put('ALLOW_TRACK_SPLITTING', true);
settings.trackerSettings.put('ALLOW_TRACK_MERGING', true);
   
% Configure track analyzers - Later on we want to filter out tracks 
% based on their displacement, so we need to state that we want 
% track displacement to be calculated. By default, out of the GUI, 
% not features are calculated. 
   
% The displacement feature is provided by the TrackDurationAnalyzer.
settings.addTrackAnalyzer(fiji.plugin.trackmate.features.track.TrackDurationAnalyzer())
   
% Configure track filters - We want to get rid of the two immobile spots at 
% the bottom right of the image. Track displacement must be above 10 pixels.
filter2 = fiji.plugin.trackmate.features.FeatureFilter('TRACK_DISPLACEMENT', 10.0, true);
settings.addTrackFilter(filter2)
   
   
%-------------------
% Instantiate plugin
%-------------------
   
trackmate = fiji.plugin.trackmate.TrackMate(model, settings);
      
%--------
% Process
%--------
   
ok = trackmate.checkInput();
if ~ok
    display(trackmate.getErrorMessage())
end

ok = trackmate.process();
if ~ok
    display(trackmate.getErrorMessage())
end
      
%----------------
% Display results
%----------------
    
selectionModel = fiji.plugin.trackmate.SelectionModel(model);
displayer =  fiji.plugin.trackmate.visualization.hyperstack.HyperStackDisplayer(model, selectionModel, imp);
displayer.render()
displayer.refresh()
   
% Echo results
display(model.toString())


Using imports in MATLAB

MATLAB lets you import java names into its workspace, using the import command. Check this. So we could rewrite the above script as follow:


%----------------------------------
% Import Fiji and TrackMate classes
%----------------------------------

import java.util.HashMap
import ij.*
import fiji.plugin.trackmate.*
import fiji.plugin.trackmate.detection.*
import fiji.plugin.trackmate.features.*
import fiji.plugin.trackmate.features.track.*
import fiji.plugin.trackmate.tracking.*
import fiji.plugin.trackmate.visualization.hyperstack.*


%--------------------
% Pick a source image
%--------------------

% Get currently selected image
% imp = IJ.openImage('https://fiji.sc/samples/FakeTracks.tif')
imp = ImagePlus('/Users/tinevez/Desktop/Data/FakeTracks.tif');
imp.show()
    
    
%----------------------------
% Create the model object now
%----------------------------
    
% Some of the parameters we configure below need to have
% a reference to the model at creation. So we create an
% empty model now.
model = Model();
    
% Send all messages to ImageJ log window.
model.setLogger(Logger.IJ_LOGGER)
       
%------------------------
% Prepare settings object
%------------------------
       
settings = Settings();
settings.setFrom(imp)
       
% Configure detector - We use a java map
settings.detectorFactory = LogDetectorFactory();
map = HashMap();
map.put('DO_SUBPIXEL_LOCALIZATION', true);
map.put('RADIUS', 2.5);
map.put('TARGET_CHANNEL', 1);
map.put('THRESHOLD', 0);
map.put('DO_MEDIAN_FILTERING', false);
settings.detectorSettings = map;
    
% Configure spot filters - Classical filter on quality
filter1 = FeatureFilter('QUALITY', 1.0, true);
settings.addSpotFilter(filter1)
     
% Configure tracker - We want to allow splits and fusions
settings.trackerFactory  = fiji.plugin.trackmate.tracking.sparselap.SparseLAPTrackerFactory();
settings.trackerSettings = LAPUtils.getDefaultLAPSettingsMap(); % almost good enough
settings.trackerSettings.put('ALLOW_TRACK_SPLITTING', true);
settings.trackerSettings.put('ALLOW_TRACK_MERGING', true);
    
% Configure track analyzers - Later on we want to filter out tracks 
% based on their displacement, so we need to state that we want 
% track displacement to be calculated. By default, out of the GUI, 
% not features are calculated. 
    
% The displacement feature is provided by the TrackDurationAnalyzer.
settings.addTrackAnalyzer(TrackDurationAnalyzer())
    
% Configure track filters - We want to get rid of the two immobile spots at 
% the bottom right of the image. Track displacement must be above 10 pixels.
filter2 = FeatureFilter('TRACK_DISPLACEMENT', 10.0, true);
settings.addTrackFilter(filter2)
    
    
%-------------------
% Instantiate plugin
%-------------------
    
trackmate = TrackMate(model, settings);
       
%--------
% Process
%--------
    
ok = trackmate.checkInput();
if ~ok
    display(trackmate.getErrorMessage())
end
 
ok = trackmate.process();
if ~ok
    display(trackmate.getErrorMessage())
end
       
%----------------
% Display results
%----------------
     
selectionModel = SelectionModel(model);
displayer = HyperStackDisplayer(model, selectionModel, imp);
displayer.render()
displayer.refresh()
    
% Echo results
display(model.toString())

Why is this greater than it seems?

Because your just ran a MATLAB script that uses and benefits from multithreading without relying on any toolbox! Here is how to tune the number of threads used by TrackMate:


% ... Do initialization before.


trackmate = TrackMate(model, settings);
trackmate.setNumThreads(3); % As many threads as you want.


JeanYvesTinevez (talk) 13:44, 17 January 2017 (CST)