Home › Forums › Software Development › [Solved] I wonder the eyeX Gaze Trace Function’s detail implementation. › Reply To: [Solved] I wonder the eyeX Gaze Trace Function’s detail implementation.

Hi!
0. Regarding gaze overlay. Yes, It’s not open-source at the moment. The tool is using a number of tricks to make it looking smooth. As you can see it’s not drawing a perfect circle but rather a morphing “bubble” instead. That helps to compensate for small hysteresis the data still has after filtering (heavier filtering will result in higher latency). But the initial data that i feed to the graphic shader is using the filter i described in a post above.
1. Everything depends on a use case.
Raw/lightly filtered data has the lowest latency so it is very useful when you design your own algorithms for object selection.
Fixation stream might be useful when it’s important to know when your eyes are doing a secade.
Interactors might be a good choice for applications related to object selection in 2d.
2. Here’s what the agreement says:
3.4 Limitations.
3.4.1 Not for analytical use. Unless Licensee enters into a separate
agreement with Tobii, the SDK may not be used to develop and
distribute software for analytical purposes, post processing or
aggregation of eye tracking data across multiple users.
Specifically, this means that such software may not (i) transfer
data about eye movements, behavior, intention or attention
generated by or derived from the Software Components to
another computing device, network, display or printer; or (ii)
store data about eye movements, behavior, intention or
attention generated by or derived from the Software
Components on any persistent storage media (except for
calibration files and data used for accuracy-enhancing filters),
i.e. data may only be stored in primary memory and only while
it is being processed.