Hi Xavier,
Thanks for your responses so far. We have a few more questions:

1. Camera focus and placement: we're able to get good focus on the face when the eyes are relatively close to the camera (60cm) but the iris and pupil images showed blurring. At a distance of ~90-100cm, the focus seems a lot better. Is this consistent with your piloting of the setup?
2. In the ReadMe there's mentioned of "Optimization/Canvas and video resolutions" but the link doesn't seem to lead anywhere. I wondered whether that info might be available somewhere.
3. We're trying out the setup with a variety of people with different eye colors/eye make-up, and wondered whether you could provide any more information on how each of the four slider setting affects image segmentation/identification of the relevant eye elements/pupil diameter.
4. Very often, we're able to get good calibration, and good data from one eye but not from both simultaneously. Initially it seemed always to be the eye closest to the red light that was well calibrated, so we bought an extra lamp. However, even with two lamps, we continue to be able to calibrate just one eye well. Mostly the problem begins with iris segmentation--the window is shifted towards the outer edge of the participant's eye. Is there anything we might be overlooking that could be causing this?
   Thanks!