The first step in linuxtrack configuration is the selection of the tracking device - just select the device of your choice from the "Tracking device" combobox.
If you can't find the device you want to use, press the "Refresh" button and check again.

Next thing that you can set is the camera orientation - normally you should not need to change this...
However, if your tracking device is in a position other that in front of you with top pointing up (you mounted TrackIR upside down for some reason, or you use laptop with a webcam chip mounted upside down,...), change the Camera Orientation to match your device's orientation.

Device Index:

Wiimote Setup
Webcam Setup
Webcam Setup for face tracking
TrackIR Setup

Wiimote Setup

If Wiimote is the tracking device of your choice, first of all, make sure you have the Wiimote server running (comes along with linuxtrack) and connected to the Wiimote.

If Wiimote server is not running, then start it, press the Connect button and press buttons "1" and "2" on the Wiimote simultaneously. After a short pause, you should see the state change to Connected and one of LEDs on your Wiimote should blink briefly about every 5 seconds.

Due to the nature of Wiimote there is no way to tweak any parameters except which LEDs should indicate running or paused tracker. Just select which LEDs should be on in the Running state and which should be on in the Paused state.

If the battery life is crucial, I'd suggest to turn all LEDs off at least in the Running state (in this state you are going to spend most of the time).

To check, how the tracking works, go to the Tracking window, select Camera view pane and press Start button to commence tracking. Ideally, you should see three (or one, in case of single point model) "blobs" - these are light emitting or reflecting points of your model as seen by the Wiimote. Each of those "blobs" should contain a white cross marking it valid.

If there are unwanted blobs, the only possibility to get rid of them is to remove them physically (e.g. turn off the lamp or cover the mirror reflecting it), since there is no way to alter the processing taking place inside the Wiimote.

Webcam Setup

To configure a webcam, at first you have to set the Pixel Format. The preferred format is YUYV, but you can experiment to see which one works the best for you.

Then continue by selecting the desired Resolution & Framerate.
The safest bet would be something around 320x240@30; when tracking on this setting works, you can safely experiment with different resolutions and framerates.

To check, how the tracking works, go to the Tracking window, select Camera view pane and press Start button to commence tracking. Ideally, you should see three (or one, in case of single point model) "blobs" - these are light emitting or reflecting points of your model seen by the camera. Each of those "blobs" should contain a white cross marking it valid.

If the webcam picture is upside down (e.g. some notebooks have the sensor mounted that way), just check the Webcam mounted upside down button.

When you try tracking for the first time, most probably you'll see more or less blobs, or white crosses that aren't in the right blobs.

Then you can try some of these solutions:

The preferred solution is the first one, although tweaking camera parameters (brightness, contrast, exposure, ...) can considerably improve the tracking.

When you have removed as much interfering light sources as possible, you might try to get rid of the rest of interfering light sources by means of setting blob size constraints.
However, this is going to work only if the interfering sources are considerable larger or smaller than "wanted" blobs.

These limits are used to discriminate false "returns" by applying a constraints on a number of pixels that form a blob.
If a blob should be valid, but it still doesn't contain the white cross, try lowering the Min. blob value and increasing the Max. blob value. It is best to fine-tune the Min. blob value when being farthest away from the camera, while the Max. blob value while being closest to the camera. When done, you should see white crosses inside the right blobs all the time, no matter how much and in which direction you turn your head (provided the webcam can still physically see the light source).

Troubleshooting tips

Webcam Setup for face tracking

To configure a webcam, at first you have to set the Pixel Format. The preferred format is YUYV, but you can experiment to see which one works the best for you.

Then continue by selecting the desired Resolution & Framerate.
The safest bet would be something around 320x240@30; when tracking on this setting works, you can safely experiment with different resolutions and framerates.

Last thing to set before first test is the path to the cascade used to track the face. OpenCV Haar and LBP cascades are supported - just browse to the cascade of your choice (normaly you should find them in /usr/share/doc/opencv or /usr/share/doc/opencv-doc; otherwise download them from web, e.g. http://alereimondo.no-ip.org/OpenCV/34). It should be a frontal face detection cascade; the most referenced cascade is haarcascade_frontalface_alt2.xml and it should provide the best results. If you are short on CPU power, try the lbp cascade lbpcascade_frontalface.xml - it consumes much less CPU, but the tracking is said to be a bit less reliable.

To check, how the tracking works, go to the Tracking window, select Camera view pane and press Start button to commence tracking. Ideally, you should see a cross in the middle of your face, that is tracking movement of your face. Remember, that in order to have as realable tracking as possible, the camera must provide good enough picture and your face should be well lit. Also a plain, single color background is preferable.

Next thing to tweak is the optimization level; it sets a ratio between the quality and speed of tracking. Try setting higher optimization level in case you want/need to reduce the CPU load caused by the tracker.

Finally, there is a Smoothing slider. If the cross is jumping wildly, you can try some of these solutions:

To steady the tracking, try to play with the Smoothing slider - moving the slider right steadies the tracking, but it also introduces a lag. Try playing with the Filter Factor on the Tracking setup tab too - these two filters have somewhat different characteristics. Exp. filter actually averages input values, smoothing the jumps, but it also introduces a lag. The filter in the Tracking setup tab helps smoothing small jitter. Try finding some sweet spot that works for you.

Troubleshooting tips

TrackIR Setup

If you are a Mac user, you can safely skip over this paragraph... When you intend to use TrackIR on linux, most probably you'll need to get access rights to the device. The easiest way to do that, is to install the 51-TIR.rules file (comes with linuxtrack) to the udev rules directory (on Ubuntu it is /lib/udev/rules, but other distros might differ a bit in this respect). When the rule is there, just re-plug the TrackIR and you should be able to access it.

If you didn't used TrackIR before, you are going to need to install the firmware. Just press the Install Firmware button and follow the instructions.

Usually just pressing Download button on the presented dialog is enough - it downloads the driver package from NP and extracts firmware needed to run the device.

To check, how the TrackIR works, go to the Tracking window, select Camera view pane and press Start button to commence tracking. Ideally, you should see three (or one, in case of single point model) "blobs" - these are light emitting or reflecting points of your model seen by the camera. Each of those "blobs" should contain a white cross marking it valid.

You can tweak the tracker state signalization and in case of TrackIR5 also the LEDs brightnesses (current limitation is, that the change will become visible upon next state change - pausing/restarting the tracking).

In case you use reflective model, you can also set IR illumination LEDs brightness.

Troubleshooting tips

If you experience unwanted reflections (from glasses, ...) while using TrackIR in the active regime (using reflective Track Clip), you might want to consider using active model with IR LEDs (like TrackClip Pro, or a home-made one).

If you see unwanted blobs, you can try one of these solutions.

The preferred solution is the first one, since no amount of processing can give you bulletproof results.

When you have removed as much interfering light sources as possible, you might try to get rid of the rest of interfering light sources by means of setting blob size constraints.
However, this is going to work if the interfering sources are considerable larger or smaller than "wanted" blobs.

These limits are used to discriminate false "returns" by applying a constraints on a number of pixels that form a blob. If a blob should be valid, but it still doesn't contain the white cross, then try lowering the Min. blob value and increasing the Max. blob value. It is best to fine-tune the Min. blob value when being farthest away from the camera, while the Max. blob value when being closest to the camera.
When done, you should see white crosses inside the right blobs all the time, no matter how much and in which direction you turn your head (provided the TrackIR can still physically see the light source).