View5D Tracking Tutorial

The image data shows DiI molecules in a lipid GUV membrane (DLPC/Cholesterol/DiI = 50/50/1e-4) at room temperature, acquired by Stefan Schäfer in the group of Petra Schwille, TU Dresden.


please wait for the viewer and the image data to be loaded (THIS MIGHT TAKE SOME TIME)! The tutorial continues down below the applet.

  alt="Your browser understands the <APPLET> tag but isn't running the applet, for some reason." Your browser is completely ignoring the <APPLET> tag!

Preparing the viewer

Since the particles or cells whose movements shall be tracked by the viewer often change their brightness (e.g. because they moved out of focus) it is often advantageous to use a non-linear color map (e.g. "glow red", as accessible by typing "c" on the keyboard) or switching on the logarithmic display mode (via "O", remember to press "shift-o"!). It is also recommended to raise ("1") the lower displayed threshold to such a level that the particle visibility is increased (here press 17 times the number one "1").

The viewer as it should look after adjustments have been performedUsing Projections

In the example given above the Z-direction has been misused as a time direction. For 2D data this way of loading the data into the viewer has advantages: Using the mode of 3D-projection, the traces of the particle x-position and y-position appear as blurry curves which offers a convenient way to check the quality of the trace. To activate the projection mode, point the mouse into the XZ- (lower right) or the YZ-view (top right) and type "p", which turns on the maximum projection. Alternatively you can use "P" for average projection mode. Both modes are toggled on and off by pressing the "p" or "P" a second time. You should now setup the above viewer to look approximately like shown on the right (click on it to enlarge!).

Trace of the track overlaid on the image.A First Track

Now in the first frame (the top frame) the brightest particle to track should be selected. Do this by placing clicking next to it and pressing "m" on the keyboard. This sets a marker. Now initiate the tracking by pressing "w". You will see a track appearing. This track will be assigned a random color, which can be changed by pressing "W" until the user likes it.
When you step through the individual planes by pressing page-up or page-down, you will notice that the track did not really follow the particle. E.g. on the second slice it should look like shown on the left. As you see the tracking has failed, even though it is OK for the next couple of slices. Setup menu for tracking including the search-parameter set to 5.0 in XY.The reason for this failure is that the particle moved between the frames further the the maximal distance the algorithm is set up to look for it.

Correct Setup for Tracking

We can change this setup by pressing "n", which will bring up the menu shown on the right (click to enlarge). As also shown on the right, change the parameters labeled X-Neighbors and Y-Neighbors to 5.0. The meaning is that for future tracking attempts a neighborhood of +/- 5 pixels along X and Y will be searched for a local maximum. After finding the maximum this procedure will be iteratively repeated until no different maximum can be found (the markers "slide uphill"). In a final step the marker position is corrected to sub pixel accuracy by calculating the center of mass in a region as defined by "X Centre of Intensity" and the appropriate Y and Z values. Also for the center of intensity calculation, the region extends from the central pixel into the + and - direction (thus 2*dx+1 pixels in size along x and so on). After changing the values, close the menu by clicking "OK" (if not visible, enlarge the window by dragging at the corner!). Return to the first slicing position by pressing the "page-up" key and reinitiate the tracking by again pressing "w".
Track after correction of the setup. Track is still too long.The track should now look like shown on the left. Note that it changed dramatically, since now the particle was followed correctly though its movement. However, the indicated slicing position as shown, is the last position at which the particle is visible. In further slices, it is lost (bleached, or moved out of focus) and the tracker tracked randomly in the noise until another particle was found. Note also that the side maximum projection shows the quality of the track as well as where the particle was lost. To minimize distortion from other particles in the projection a user-defined region of interest (ROI) was selected (by first pressing "S" and then shift-left-mouse-dragging and then clicking on the individual corners and double-clicking to close the ROI). Note that the ROI can be deleted by a shift-double-click with the left mouse button.
To now eliminate the remainder of the track, browse to the first position where the particle is not visible any more and press "Q". This will delete all the remaining markers, including the active marker. This track is now finished.

Tracking Multiple Particles

To track a second particle, go back to the first slice, where this particle is visible (not necessary the first slice of the stack), point the cursor onto the particle (left click) and press "k". You will see how another marker, not connected to the other markers appears. Markers are organized in marker lists. "k" creates a new marker list. "K" deletes the active marker list. With "j" and "J" the user can browse through the existing marker lists. For more information on working with markers and marker lists click here.

All tracks are completed. One marker position was deleted and one changed by hand for the track on the lower right.Deleting Track Position

Play with the viewer and try to track all the particles that are visible. The result should look like shown on the right. You will notice that the particle on the bottom right was not tracked correctly at the plane (#9) as shown on the right. If you are not sure where it is (e.g. due to single molecule blinking) simply delete the marker at this slice by pressing "M".

Checking the Track

Another convenient way to browse though tracked lists is pressing the number keys "0" and "9". In contrast to ordinary browsing via "page-up" and "page-down", this approach always updates all slicing positions to match with the marker. Thus it is immediately possible to see whether the marker is still "on-target". When tracking 3D-datasets over time this feature is essential and makes work a lot easier!

Helping Out and Tracking "by Hand"

If the tracking failed, you can try to "help" the algorithm by simply dragging the marker close to where you think the particle was. In the mentioned case this is possible, but you have to drag it to a position that the algorithm does not find some other noise pixel. However, it is also possible to temporarily toggle off the "use automatic maximum finding" part of the algorithm in the menu accessible via "n" (see also menu panel above). Once this is turned off, the marker position after a drag is only influence by the local center of mass, which can also be turned of if needed. Finish the track until you find them correct.

Tracking in Multi-Color images

If multi-color images are used for tracking, it is recommended to switch into single-color mode ("C") whilst doing the tracking! The reason is that always the active element will be used for tracking and in multi-color mode the user is often not aware that a different than the intended element is "active" (red border around element color map). This then leads to seemingly wrong tracking results.

Saving the Result

When the tracks are finished, the result should be saved. This is achieved by pressing "m" in the lower right part of the window (the element display). You will see a textual display in the window. From all the text can be marked (left-drag) with the mouse via the system dependent copy mechanism (ctrl-c for windows systems) and pasted into a spread-sheet program like Microsoft Excel or any text editor. However the program also tries to directly save the results to disc. However, when running as an applet this is often not possible due to security restrictions. In some systems the security can be adjusted to allow saving data from java programs. The name and path under which the program tries to save can be entered in the setup menu accessible via ("n"). Click here for further hints on influencing the java security. Under Matlab it is possible to directly obtain the marker information as a list of Matlab vectors. For details click here.

Loading Markers

Similarly Lists of marker positions can be loaded and displayed (press "L" in one of the main displays). This will lead to an update of the positions of existing markers according to their positions in the file and new markers will be created if necessary. However, no markers or lists of markers will be deleted.
In Matlab it is also possible to add markers to an existing viewer instance (click here). The marker file names can also be provided via the applet tags for the applet version.

alt="Your browser understands the <APPLET> tag but isn't running the applet, for some reason." Your browser is completely ignoring the <APPLET> tag!Preprocessing

The program View5D is not constructed for performing major data processing. However, depending on your application, it may be very advantageous to do so. You can for example pre-process the data in ImageJ or Matlab before you start View5D to do the tracking. A simple filtering operation can yield a different view of the data (see viewer with tracks on the left, and browse though them).

Tracking direction

Tracking can be performed along any of the directions. The chosen tracking directions is configured via the menu accessible via "n".





















Back to the View5D homepage
For hints and suggestions, contact the author under heintzmann at gmail dot com