TrackMate - incorrect nSpots in xml file

[Daniel Matthews]

Hi

I have recently started using the fantastic TrackMate plugin for FIJI. I am doing post tracking analysis in Matlab and so am trying to parse the data with the "importTrackMateTracks" script. After running the LAP tracker with splitting and merging I am finding that the xml results isn't always parsed and the reason seems to be that the "nSpots" attribute isn't being written correcly. For example, I have a node where nSpots="90" but there are are only 86 detections and so the "importTrackMateTracks" script throws an error.

Cheers,

Dan.

[Daniel Matthews]

For those interested, a workaround to the problem is to amend "importTrackMateTracks" as follows:

        if detectionNodes.getLength ~= nSpots
            nSpots = detectionNodes.getLength;
        end
       
        A = NaN( nSpots, 4); % T, X, Y, Z

Which determines the real number of detections in the node, and if "nSpots" differs from this value it gets reassigned. You then need to initialise A after doing this.

Cheers,

Dan.

[Jean-Yves Tinevez]

Hi Daniel,
This is annoying. Any chance you could send me via email the faulty XML file?
Coupled to the recent imglib2 update, there will be a TrackMate release which I would like to focus on bug fixes.
Best
jy

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html