Generally when tracking eye movement we use various methods that require sensors being pointed at the eye itself. This approach is quite different in that it is sensing the “electrical potential of the cornea”. We have no idea how this works, but it looks pretty cool.
[via HackedGadgets]
http://en.wikipedia.org/wiki/Electrooculogram
It’s a little surprising that optical techniques for eye tracking are so much more popular.
interesting. RObert Zemeckis (back to the future) used this to record the actors eye movements in Beowulf for accurate translation to the digital models.
My friend did a similar thing for his senior project 2 years ago, with a fairly substantial write-up: http://jotux.com/projects/senior-project-eyemouse/
I don’t know if it’s still the method of choice, but in the early days of sleep research (late 60’s, early 70’s) this same principle was used to detect REM sleep. The basic idea is that because the eye is an electrochemical device with all the “wires” (nerves) running out the back, the eye has an electrical potential. Here’s an article on Wikipedia: http://en.wikipedia.org/wiki/electro_oculography
@Kevin H
Thank you for the wonderful article. I found it much more informative then the one actually featured. :) I only wish he took it further. I would have loved to see him continue to develop this into a working control system. Anyway, cool link. Thanks.
that looks funny…..
i wonder when we’ll get a DIY device which points your eyes in a certain direction :O
ps: no, pr0n does not count :-/
Cool!! Very interesting.
I’ve done something similar. It’s easy and cheap although you can have some precision problems caused by the electrodes contact. Yo must clean your skin well and/or make something to hold the electrodes in place.
The problem with EOG interfaces is that the position is basically a small DC signal, which tends to saturate any amps. If you highpass at .1Hz or so, you can get some decent tracking, but it’s only accurate to a few degrees, which can translate to a significant distance on a monitor. Plus, if you filter DC out, then you can only tell when you move, since your signal will go in one direction, and then level out back to zero. You’d use that waveform’s magnitude to figure out how far your eye moved.
There’s a reason we don’t see EOG equipment around very much. Still cool stuff though.