Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Supported by

[solved] Data not recorded properly (PyGaze, EyeTribe)

edited May 2016 in PyGaze

Hi everyone,

I made an experiment where one has to read a small text in OpenSesame, using PyGaze to record eye movement from the EyeTribe.
The problem is: I have a huge loss of data in the resulting .tsv file, as if the eyes were closed (lines of zeroes). Also, if the EyeTribe samplerate is set to 60Hz, the data of almost half the text are not recorded. At 30Hz, this last problem doesn't seem to occur though.

I thought it might be an issue with OpenSesame, so I made a script directly in Python with PyGaze. I still have the same problem...
If it can help, I made sure that the average calibration error was always smaller than 1, I used multiple computers (but only one EyeTribe).

Some images of the scanpaths created with the resulting .tsv:

At 30Hz with PyGaze:

image
And the associated .tsv here

At 60Hz with OpenSesame:

image
And the asociated .tsv here

At 30Hz with OpenSesame:

image
And the associated .tsv here

Cheers,

Comments

  • edited 12:52PM

    Hi,

    Like you said yourself; "as if the eyes were closed". This is correct: one has to blink occassionally. During a blink the coordinations can go all over the place, due to loss of signal or a misinterpretation (e.g. eyelashes). On top of that, the eyetribe doesn't have a great spatial and temporal resolution (although of course great for its price!). What you could at least do to get decent scanpaths, is to make sure only fixations within the text area are logged.

    Hope this helps.

    Cheers,

    Josh

  • edited 12:52PM

    Hi Josh,

    Thanks for your reply. I am well aware that one has to blink and that it can cause a loss of the signal. Thus, I didn't think the loss of signal I encountered was due to blinking. There are a lot of them that last more than a second. Also, I have done some testing in different conditions, making sure I and the other testers were not blinking too much but the results are still the same.

    I tested a script I made from an example that draws a dot "following" the gaze in real-time and I don't have these patterns of signal loss. That is why I thought it was an issue occuring during the logging.

    I used PyGaze analyser to parse the tsv files and to draw the scanpaths. From what I unsterdand, PyGaze analyser automatically detects blinking and do not use them for the scanpaths. So I shouldn't have fixations drawn at (0,0).

    I will take a closer look into this, thank you.

    Cheers,

  • edited 12:52PM

    Hi,

    Essentially as @josh said: 0,0 coordinates indicate signal loss, which can occur (and usually does occur) for any number of reasons. The solution is simply to remove or ignore all 0,0 coordinates.

    Cheers,
    Sebastiaan

    There's much bigger issues in the world, I know. But I first have to take care of the world I know.
    cogsci.nl/smathot

  • edited 12:52PM

    Hi,

    I have modified the PyGaze script to ignore the (0,0) coordinates so I can use the data for my analysis. Still, The loss of data is quite an issue as nearly half of the recordings is lost. I think it might be a problem with my EyeTribe so I will try to find another one to test it out.

    Thank you both,
    Cheers,

Sign In or Register to comment.