Parameter details

• Motion models are selected based on the captured motion. Generally Constant speed works well, however if you are not sure about the selection, check all of the models and the algorithm will evaluate the best fitting model by itself.

• In order to speed up the tracking it is a good idea to turn off all Object Features and Track Processing features for few initial tries.

• The Max Object Speed (maximum cut off speed) should be set first because it restricts the number of combinations the tracking algorithm must try. It is an absolute value which eliminates impossible links right away and no two objects can link if the speed is greater than this limit. Usually, the maximum speed of observed objects is well known for the investigated phenomenon. If you know the exact value, type it into the white field. If the object speed limit is not know, use the graph below to find out the speed value under which most of the tracks appear.

• The goal of the initial setup is to find the Standard Deviation Multiplication Factor for the given sample while reducing the Maximum Object Speed as much as possible.

• Standard Deviation Multiplication Factor (tracking rigidity) is a relative to each object and its history and represents the probability of non-linking. It cuts when the motion would be unexpected. There is an aid to understand it better - the Show probability ellipse in the Display tab. It works after first tracking has been made. A small cross with a solid ellipse is shown for each track point. The cross represents the predicted object position. If an object is found within the predicted restriction limit represented by the circle, its position becomes part of the resulting track. By changing the values, the ellipse grows or shrinks. Dashed ellipse is drawn around predicted point where the object was not found. It should be set so that all consecutive points are contained in the corresponding ellipses.

• Move the sliders in the Object Tracking tab to see the changes. The goal is to make the circles a little smaller to make gaps and then close them.

• If the sample is too dense and sampling in time sparse there may be too many points in each ellipse making the probability of failing higher (objects are assigned to incorrect tracks). In this situation it may be better to start the experiment with the Close Gaps Between Tracks function.

• When you have made the probability ellipses a bit smaller, your objects may sometimes disappear and therefore cannot be tracked in these frames. You should reconnect the separated tracks back to the original path by the Close gaps between Tracks function. In general as the Maximum Gap Size gets higher the prediction accuracy deteriorates quickly.

• When the rough combination of Standard Deviation Multiplication Factor, Max Object Speed and Max Gap Size is determined it is wise to switch on Allow New Tracks After First Frame and use functions in the Track Processing tab to filter out some bad tracks.

• If objects that belong to the same track have some distinguishing features (e.g. if there are large and small objects) switch on EqDiameter or other features from the Object Features tab to see if it improves the result.

• When objects cannot be segmented nicely and they break occasionally into smaller pieces it is good to break the tracks in such a place, let introduce a gap and then close it in Track processing. This requires finding the correct limit for EqDiameter or Circularity. The limit is entered as percentage change per second which may yield weird numbers depending on the timelapse frequency.

• Export your data and tracks to MS Excel or Clipboard, and save graphs as raster files or copy them to the clipboard. Mean square displacement (Mean square displacement (MSD)) of the tracks can also be exported to MS Excel and used for the Intra-Nuclear Single Particle Tracking (Intra-Nuclear Single Particle Tracking (I-SPT)) to indicate the type of diffusion.