The Eye Tribe

Hello all,

I've recently started playing with my TET and I have some questions that I either can't seem to answer using the website, or due to my technical knowledge.
1. Just curious, what algorithm is TET using to classify fixations? I-DT?
2. Is pupil size measured in millimeters or pixels?
3. I've been unable to find any documentation on what is done to position data that is "smoothed." Can someone please clarify what type of smoothing that is applied to the position data, and how this is executed?
4. Forgive my ignorance, because I know this is mentioned on the website, but I'm not familiar with 32-bit masked state integers. Is it possible to list the possible values that TET outputs, and what they mean? For example, I've seen state codes of 4, 7, 8, and 16 in the short time I've started exploring.

Thanks in advance for any support anyone may be able to provide. Again, forgive any stupid questions, I haven't yet programmed in C++ or C#, so some of this is quite new to me.

Bill

Welcome to the community

The algorithm to detect fixations cannot be disclosed, but it uses dispersion information. Once a fixation is detected, the signal is smoothed to reduce jitter. Again, we are not allowed to disclose how this is implemented. Please note that the API also provides the raw gaze data, so you are very welcome to develop your own fixation detection algorithms.

Pupil size is currently measured in arbitrary units. In a future release we will provide the size of the pupil in millimeters.

The masked integers are used to represent the state, and can take many values for a given sample. You can take a look at the C# SDK and samples on GitHub to see how you can check whether a specific state is "present" in the integer received over the API. In particular, check the ProcessSample method in the TrackBox sample.

WDMiller wrote:4. Forgive my ignorance, because I know this is mentioned on the website, but I'm not familiar with 32-bit masked state integers. Is it possible to list the possible values that TET outputs, and what they mean? For example, I've seen state codes of 4, 7, 8, and 16 in the short time I've started exploring.

Javier wrote:The masked integers are used to represent the state, and can take many values for a given sample. You can take a look at the C# SDK and samples on GitHub to see how you can check whether a specific state is "present" in the integer received over the API. In particular, check the ProcessSample method in the TrackBox sample.

I have the same question and problem. I would like to use the "state" column of the output files to filter for blinks and data loss in general. I'm not familiar with programming in C or C# either and the 'ProcessSample' code that Javier pointed to did not help me with the interpretation of these values.

Support requests in this matter are always answered with a reference to look at the API, yet all I could find is the following:

Tracker State (http://dev.theeyetribe.com/api/)
When requesting the tracker device connectivity state value trackerstate it returns an integer/enum with a value as described below:
State Description Value
TRACKER_CONNECTED Tracker device is detected and working 0
TRACKER_NOT_CONNECTED Tracker device is not detected 1
TRACKER_CONNECTED_BADFW Tracker device is detected but not working due to wrong/unsupported firmware 2
TRACKER_CONNECTED_NOUSB3 Tracker device is detected but not working due to unsupported USB host 3
TRACKER_CONNECTED_NOSTREAM Tracker device is detected but not working due to no stream could be received 4

While it appears to me that 7 indicates successful tracking of both eyes, 4 indicates partial or complete loss of tracking information (one or both eyes). 8 is coupled to all values being 0 and a sample with state 8 often precedes several samples with state 4, i.e. data loss. So my guess would be that 8 indicates a more severe tracking problem, for example that in addition to the eyes the head is not detected/located. Haven't encountered a 16 or other values, yet.

Would it be possible to just get a list of possible values and their meaning? Or am I missing the complexity of these values?

Many thanks!