Howdy, Stranger!

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

Supported by

eye tracking from previously recorded IR illuminated video

Are there any tools to identify glint/pupil and provide eye tracking for a prerecorded eye video?

We have had a hard time getting reasonable pupil discrimination (Arrington's viewpoint) likely because of some pretty serious noise. Unfortunately, parsing eye movements is necessary to scoring the already completed fMRI task. Fortunately, we've recorded video.

If we can clean up the video after the fact, are there any tools we can use to extract estimated gaze?

I'm digging into pypupil/pupil-labs source now, but am hoping someone has already done the hard work.

noise gif

Comments

  • Ooph, that's a pretty noisy video indeed! I don't have a lot of my code out publicly, but I do this type of thing quite frequently. I'd be happy to collaborate.

    Drop me a line at edwin.dalmaijer@mrc-cbu.cam.ac.uk , if you're interested. We can try a single example video first, to see whether my tools work.

Sign In or Register to comment.