(Simple) LAP Tracker: use Jonker-Volgenant instead of Munkres-Kuhn

Simple tests showed that Jonker-Volgenant has nicer expected runtime than Munkres-Kuhn, important for our rather large problems... Signed-off-by: Johannes Schindelin <johannes.schindelin@gmx.de>
trackmate-sc · Jan 22, 2014 · ad50128 · ad50128
1 parent 4f0acc3
commit ad50128
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 3 deletions.
diff --git a/src/main/java/fiji/plugin/trackmate/tracking/FastLAPTracker.java b/src/main/java/fiji/plugin/trackmate/tracking/FastLAPTracker.java
@@ -4,7 +4,7 @@
 
 import fiji.plugin.trackmate.SpotCollection;
 import fiji.plugin.trackmate.tracking.hungarian.AssignmentAlgorithm;
-import fiji.plugin.trackmate.tracking.hungarian.MunkresKuhnAlgorithm;
+import fiji.plugin.trackmate.tracking.hungarian.JonkerVolgenantAlgorithm;
 
 public class FastLAPTracker extends LAPTracker {
 
@@ -15,6 +15,6 @@ public FastLAPTracker( final SpotCollection spots, final Map< String, Object > s
 
 	@Override
 	protected AssignmentAlgorithm createAssignmentProblemSolver() {
-		return new MunkresKuhnAlgorithm();
+		return new JonkerVolgenantAlgorithm();
 	}
 }
diff --git a/src/main/java/fiji/plugin/trackmate/tracking/FastLAPTrackerFactory.java b/src/main/java/fiji/plugin/trackmate/tracking/FastLAPTrackerFactory.java
@@ -13,7 +13,7 @@ public class FastLAPTrackerFactory extends LAPTrackerFactory
 
 	public static final String NAME = "LAP Tracker";
 
-	public static final String INFO_TEXT = "<html>" + "This tracker is based on the Linear Assignment Problem mathematical framework. <br>" + "Its implementation is adapted from the following paper: <br>" + "<i>Robust single-particle tracking in live-cell time-lapse sequences</i> - <br>" + "Jaqaman <i> et al.</i>, 2008, Nature Methods. <br>" + "<p>" + "Tracking happens in 2 steps: First spots are linked from frame to frame to <br>" + "build track segments. These track segments are investigated in a second step <br>" + "for gap-closing (missing detection), splitting and merging events.  <br> " + "<p>" + "Linking costs are proportional to the square distance between source and  <br> " + "target spots, which makes this tracker suitable for Brownian motion.  <br> " + "Penalties can be set to favor linking between spots that have similar  <br> " + "features. " + "<p>" + "Solving the LAP relies on the Munkres-Kuhn solver, <br> " + "that solves an assignment problem in O(n^3) instead of O(n^4)." + " </html>";
+	public static final String INFO_TEXT = "<html>" + "This tracker is based on the Linear Assignment Problem mathematical framework. <br>" + "Its implementation is adapted from the following paper: <br>" + "<i>Robust single-particle tracking in live-cell time-lapse sequences</i> - <br>" + "Jaqaman <i> et al.</i>, 2008, Nature Methods. <br>" + "<p>" + "Tracking happens in 2 steps: First spots are linked from frame to frame to <br>" + "build track segments. These track segments are investigated in a second step <br>" + "for gap-closing (missing detection), splitting and merging events.  <br> " + "<p>" + "Linking costs are proportional to the square distance between source and  <br> " + "target spots, which makes this tracker suitable for Brownian motion.  <br> " + "Penalties can be set to favor linking between spots that have similar  <br> " + "features. " + "<p>" + "Solving the LAP relies on the Jonker-Volgenant solver, <br> " + "that solves an assignment problem in O(n^3) instead of O(n^4)." + " </html>";
 
 	@Override
 	public String getName()