From: dstamp@watserv1.waterloo.edu (Dave Stampe-Psy+Eng)
Subject: Re: Eye-movement interface
Date: Sat, 11 Jan 1992 03:12:07 GMT
Message-ID: <1992Jan11.031207.18204@watserv1.waterloo.edu>
Organization: University of Waterloo



M22367@mwvm.mitre.org writes:

>I am ISO information regarding eye-movement tracking systems.  I know such
>systems exist to aid persons who are physically incapacitated - I need to
>find names and addresses, please.

It depends what kind of eye-movement systems you're after.  There are
3 main types:

- electro-oculoagraphs which use electrodes just beside the eyes to
sense position.  Unfortunately they're noisy, drift and can only sense
horizontal eye movements.  Plus noone is quite sure what they're sensing

- schleral reflection uses phototransistors and LEDS (infrared) mounted
on an eyeglass frame.  Pulsed light and synchronous detection is used
to reject room light.  Effectively, they sense the relative area of
schlera and iris under each phototransistor and try to predict eye
position from this.  They're cheap, but very inaccurate and hard to
calibrate.  They usually can only measure horizontal eye movements
reliably, and have a lot of crosstalk with vertical eye movements.

- imaging pupil trackers actually watch the eye with a camera, and
use a realtime video processor to extract the position of the pupil
of the eye.  As can be infwerred, they're expensive (10000$ or more)
but quite accurate.  If the eye camera is not mounted on the subject's 
head (miniature "finger" camera and headband) the subject's head must
be held very still-- you can see breathing with these things.

- corneal reflection trackers bounce an IR point source off the 
curved cornea of the eye, and can track position well (if over
a limited range).  Setup tends to be a little tricky, but either
imaging or some cheaper technique (graded response with position,
for example) may be used.  Cost depends on the measuring technique 
used.

My own work has been mostly with imaging eye trackers.  Recently I
completed a head-mounted system that used a second camera to track
LEDs mounted on a presentation monitor for head position sensing.
Errors were less than 1 degree in eye-pointing accuracy with fairly
wide head motions (most subjects can't fix their gaze nearly that
well while moving their heads, so it's hard to measure).

Sources: ISCAN Inc. makes and sells imaging pupil and corneal
reflection tracking systems, but I'm not thrilled with their
head mounts.  Biometrics makes a schleral reflection eye tracker,
and I know of several universities with them.  (I also know that
getting them to work is a black art, and the people who work with
them look with envy on imaging tracking systems).  I don't think 
anyone seriously uses EOG measureing systems anymore.  "Cheaper"
corneal reflection systems wait for the proper market to evolve,
IMHO.

--------------------------------------------------------------------------
| My life is Hardware,                    |                              | 
| my destiny is Software,                 |         Dave Stampe          |
| my CPU is Wetware...                    |                              | 
| Anybody got a SDB I can borrow?         | dstamp@watserv1.uwaterloo.ca |
__________________________________________________________________________