Trackmate - Tracking straight High Speed movements

Hey,

Same as trackmate-tracking-at-high-speed i have some issues to track High Speed particles with Trackmate.
In my analysis the particles have very straight movements with roughly constant velocity vectors:
Substack (1-200).zip (9.8 MB)

Because of that i am using the Kalman Tracker (Linear Motion Tracker).
It works well, but not perfect.

Because of the High Speed, i need a large search radius. And this leads to the fact, that Trackmate detects Tracks between one particle and another particle, which is closer than the same particle a frame later. You can see the edges here:

or as .avi: TrackMate capture of Substack.zip (19.7 MB)

Did anybody have the same issue and how did you fix it?

If the Tracking Algorithm would have the information, that adding spots to the Track only makes sense if the new spot is in line (Trendline of the Track would intersect the new spot) to the Track.

Another way to do it would be with the angle from one spot to the following spot of a Track. If the angle is above a value the spot will not be added.

Is it possible to modify the Tracking Algorithm?

Thank you!
Thomas

1 Like

It’s certainly possible to write your own (or adjust an existing) tracking algorithm:

This will require some Java coding though; but the extensive tutorials by @tinevez are really helpful and I highly recommend them.

In your case, it would probably be helpful to add the source of the particle movement – the center of the masked area – as a starting point for each track. That way, the track direction would be constrained to point radially outwards always, even when the first two spots are being linked.


But let me also suggest an alternative analysis approach:

  • take the maximum intensity projection across time:

image

(btw, this also reveals some line artifacts of your detection system)

  • then you can make line selections that you can restore on the original stack:

image

  • and then reslice (Image > Stacks > Reslice…) to get a kymograph where the slope corresponds to the speed of the particle:

image