Object size

The tracking is based on computing the center of mass in a region whose size you can adjust here.
The "Object size" should be large enough to enclose one of your objects of interest, but small enough not to enclose two objects as this can cause the algorithm to 'switch' objects during tracking.

Tracking factor

The factor determines the window in which the algorithm will search for the object in the next time-point. The window size is (Object size * Tracking factor). This needs to be large enough to cover the distance that the object moves between subsequent time-points. If the objects moves out of the "Tracking window" the algorithm will not find it anymore.

Tracking sub-sampling

To increase the tracking speed you can analyse only every dx,dy,dz data point. Especially when working with Tiff stacks this will help as loading only every dz plane will be much faster. Center of mass computation is so fast that dx and dy sub-sampling do not help much. However, for the correlation-based tracking dx and dy sub-sampling makes a big difference in speed.

Track length

Here, you can specify the number of frames that you would like to track the object. Please note that you do not need to start tracking at the first frame. Simply move the time slider on the image forward and then select an object and click 'Track selected object'. Also note that you can track different objects for different lengths.

Cropping factor

The factor determines the size of cropped object view. The size is: Object size * Cropping factor.

Set nx, ny

Place a rectangluar selection on the image and click this button to read its width and height.

Set nz

Click twice at different positions of the image's z-slider to set the z range.

Track selected object

Use ImageJ's Point Selection tool to select an object in the image and then hit this button to start its tracking. You do not have to wait for the tracking of one object to be finished; you can immediately start tracking another object.

Show track table

Shows a table with the current tracks.
Newly computed track positions will be added on the fly.
All columns can be sorted by clicking on the column names.
The currently selected object will be highlighted in the image.

Save track table

Saves the table as a text file for further processing, e.g. in R or Excel.

Clear track table

Removes all tracks from the track table.

Show tracked objects

Opens as many new image windows as there are completed tracks. Each image window shows the data locally surrounding the respective track. The size of the shown region can be adjusted by the changing the values in 'Cropping radii'.

Main ideas behind this functionality

One often wants to observe morphological changes of moving objects. It is however very challenging for the human brain to focus on the morphological changes as it is distracted by the object's motion. Thus, having a 'stabilized view' on the object, where its translocation is eliminated is very helpful. In addition, the 'stabilized view' also reduces the amount of data needed to further study the object.

Report issue

Opens the github issue web-site.

Tracking Method

You can choose between 'center of mass' and 'correlation'. For both methods, the mean intensity within the current tracking window **is subtracted** from each pixel as a background removal strategy.

center of mass

The center of mass is very fast to compute. For 'center of mass' the position at each time-point is determined by an iterative (n=10) center-of-mass computation, where the size of the region in which the computation happens gradually shrinks from "Object size * Tracking factor" to "Object size".

correlation

For 'correlation' the position of the first time-point is determined as its center of mass. The position of subsequent time-points is determined relative to the previous time-point by computing the shift that yields the maximal cross-correlation between the signals.