[Jason S. Babcock] and [Jeff B. Pelz] put together this paper on building a simple, lightweight eyetracker (PDF) to foster the creation of open source eyetracking software. All of the components are mounted to a cheap pair of safety glasses. The eyetracker uses a technique called “dark-pupil” illumination. An IR LED is used to illuminate the eye. The pupil appears as a dark spot because it doesn’t reflect the light. A bright spot also appears on the cornea where the IR is directly reflected. An eye camera is mounted next to the IR LED to record the image of eye with these two spots. Software tracks the difference between the two spots to determine the eye orientation. A laser mounted to the frame helps with the initial calibration process. A scene camera placed above the eye records what the eye is viewing. The video from these two cameras can be compared in real time or after the experiment is concluded.
[thanks austin y.]
This is really cool, I remember someone asking for this not too long ago, I hoep he/she is still reading.
I wonder if this could be adapted to make a head tracker? Attach a strong IR source at camera to the top of the monitor, and find the bright and dark dots that appear/disappear in sync with the IR modulation to locate the eyes. That could be used to then update the display on the screen, making it appear as a window onto a 3D world (a technique known as fishtank VR).
Looks like a great idea, I can see this being used by handicap people to communicate, use computers or even play games, who knows, maybe in the future Hawking will own you in HALO 4?
Btw, I found some infrared hazard notes here: http://www.twi.co.uk/j32k/protected/band_13/faq_radiation.html
and here:
http://www.potterymaking.org/safetyeyes.html
this is really cool, and the applications for the disabled would be awesome if they were F/OSS.
We had a family friend get in an accident and he was quadriplegic, and his eye-tracking computer system was nearly $10,000.
woohoo, my first submitted hack made it!!
This reminds me of the pupil pistol from Robotech. I sure hope this doesn’t end up in the military’s hands…the pupil pistol in Robotech started out for playing games and ended up in use for the hovertanks if I’m not mistaken.
There are lots of good files if you drop back into the publications folder
http://www.cis.rit.edu/people/faculty/pelz/publications/ETRA04_rothkopf_pelz.pdf
I can see (no pun intended) an application for this type of hardware in usability studies (http://www.poynterextra.org/eyetrack2004/main.htm).
Specifically for software interfaces such as web pages.
You would want to replace the scene camera with a recording of the current screen that the user/subject is working on.
I’d be quite interested as to how users interact with the sites I work on.
This is all well and good, but without ready-made setups available for commercial purchase, this hack will stay out of the hands of many willing armchair developers. It’s like Lego Mindstorms. Sure robotics existed before, but Mindstorms put into a package most non-hardware hackers were able to deal with, and the same thing needs to happen to this.
We, the software-only/mostly hackers, need a prepackaged setup, complete with assembled or ready-to-assemble hardware and bundled software to get us started. And then, you’ll see a boom in all sorts of hobby applications!
Can’t wait to get my hands on one then!
BTW, has anyone commented on the long-term effects of constant IR to the eye?
a setup where you have this on, some cameras mounted to triangulate your head in 3-d, and a switch you bite with your teeth could let quads (like christopher reeve was) use computers easily, and since with a bit of work you can make computers do just about anything, the possibilities are endless!
There’s actually several ways to track an eye, this is probably one of the easiest. Before cheap and plentifule mini cameras, they did it with two IR LEDs and two IR photodiodes mounted in pairs on the sides and top and bottom of the eye. These were pulsed and the alternating photodiode readings taken to get an IR intensity scale. Fairly accurate IIRC, 1-2 degrees.
Another popular way is to directly EOG the eyes. Place either passive or active electrodes on the bridge of the nose, top and bottom of the eyes, the side of the eyes, and a reference behind the ear (6 electrodes total). The eyes have something like a 100mV potential front to back, so the electrodes would pick up this voltage difference as your eyes rotated. This method was possibly more accurate, but either needed stick-on electrodes or expensive silver active electrodes. Electrode placement had to be spot on and it had to be recalibrated for each user, however.
The army uses direct laser illumination of the eye for tracking in the prototype Comanche helmets. I don’t know if that tech will filter down to the Apaches now that Comanche is canceled.
Oh, and speaking of head tracking, Apache helicopters do head tracking by using two sensors mounted behind the pilots head and two IR(?) LEDs on either side of the pilot’s helmet. It’s been a while since I read that manual.
I’m personally interested in getting a rugged eye tracking system designed for portability, as I’m also slowly working on a head mounted laser or LED monochrome scanning display to go along with it. Hmmm… maybe I’ll have to go learn my VHDL and open source the hardware variant.
Great catch.
I wonder how it will help in developing some hardware that can help in diagnosing autism. I have written about this here.
(part 1)
http://nuit-blanche.blogspot.com/2005/06/eye-tracking-machines-and-autism.html
(part 2)
http://nuit-blanche.blogspot.com/2005/11/eye-tracking-machines-and-autism.html
The reason I wrote these entries is mostly because each commercial eye tracking device cost at least $20,000.
Igor.
Here is an improvement on this eye tracker by Derrick Parkhurst ( http://hcvl.hci.iastate.edu/cgi-bin/parkhurst.cgi ) at Iowa State.
http://hcvl.hci.iastate.edu/cgi-bin/openEyeswiki/index.cgi?action=search&search=openEyes
A mobile eye tracker (with software) that seems to cost about $300 without the attendant PC. Outstanding.
Igor.
i would love this, to bad i have a stigmatism
Hi all,
Trying to cost this out, and I’m coming up to $200 for the two cameras, another $12 for a laser diode, and a few cents each for an IR LED and voltage regulator, but I’m missing the “how to hook it up to the computer” bit, and also how to power it – do we need to have a rechargable battery? Powered over USB? Do we communicate over bluetooth, IR or something else?
I’d be interested in building one of these as a project and getting it working as a pointer.
Cheers,
Dave.