WEBVTT

00:00:00.000 --> 00:00:01.320 align:middle line:90%


00:00:01.320 --> 00:00:04.770 align:middle line:84%
And should we talk about
how we do calibration

00:00:04.770 --> 00:00:06.520 align:middle line:84%
with the mobile
trackers as well?

00:00:06.520 --> 00:00:07.900 align:middle line:90%
Yeah.

00:00:07.900 --> 00:00:08.680 align:middle line:90%
I can wear it.

00:00:08.680 --> 00:00:11.550 align:middle line:90%


00:00:11.550 --> 00:00:14.520 align:middle line:84%
So the process here is
a little bit different.

00:00:14.520 --> 00:00:19.720 align:middle line:84%
Since people are not usually
looking at a computer screen,

00:00:19.720 --> 00:00:22.770 align:middle line:84%
but instead out in
the world, there is--

00:00:22.770 --> 00:00:24.790 align:middle line:84%
different systems use
different processes.

00:00:24.790 --> 00:00:26.740 align:middle line:84%
But there's kind of
two general options.

00:00:26.740 --> 00:00:29.850 align:middle line:84%
One is kind of similar
to what we just

00:00:29.850 --> 00:00:31.800 align:middle line:84%
showed with this
stationary eye tracker.

00:00:31.800 --> 00:00:35.940 align:middle line:84%
That you use the software
that produces a visual target.

00:00:35.940 --> 00:00:38.670 align:middle line:84%
And that can be displayed
either on a computer screen

00:00:38.670 --> 00:00:41.280 align:middle line:84%
or on a wall or a
projector if you

00:00:41.280 --> 00:00:43.170 align:middle line:84%
want to capture more
of the visual field.

00:00:43.170 --> 00:00:47.070 align:middle line:90%
And the person stands straight.

00:00:47.070 --> 00:00:48.270 align:middle line:90%
They keep their head fixed.

00:00:48.270 --> 00:00:50.010 align:middle line:84%
And they look around
with their eyes.

00:00:50.010 --> 00:00:52.110 align:middle line:84%
And it's important
that the person

00:00:52.110 --> 00:00:55.020 align:middle line:84%
stands with their head fixed
and looks with their eyes

00:00:55.020 --> 00:00:56.250 align:middle line:90%
and not with their heads.

00:00:56.250 --> 00:00:59.100 align:middle line:84%
Otherwise, you end
up getting, actually,

00:00:59.100 --> 00:01:03.300 align:middle line:84%
a very small part of the visual
field calibrated, rather than

00:01:03.300 --> 00:01:05.280 align:middle line:84%
the full range of where
your eyes might look.

00:01:05.280 --> 00:01:09.630 align:middle line:84%
So do we have to recalibrate the
stationary eye tracker as well

00:01:09.630 --> 00:01:12.220 align:middle line:84%
as the mobile tracker when
you are running an experiment?

00:01:12.220 --> 00:01:13.710 align:middle line:84%
For example, if
something happens--

00:01:13.710 --> 00:01:16.530 align:middle line:84%
if I move the eye tracker
during an experiment,

00:01:16.530 --> 00:01:18.900 align:middle line:84%
do you have to
recalibrate, you can do it?

00:01:18.900 --> 00:01:20.820 align:middle line:84%
Usually, it's a good
idea to recalibrate

00:01:20.820 --> 00:01:23.310 align:middle line:84%
because now the cameras
are in a different position

00:01:23.310 --> 00:01:24.810 align:middle line:90%
relative to the eye.

00:01:24.810 --> 00:01:28.200 align:middle line:84%
And so probably, the calibration
is now off a little bit.

00:01:28.200 --> 00:01:31.530 align:middle line:84%
And your data will not
be quite so precise.

00:01:31.530 --> 00:01:34.380 align:middle line:84%
So for the mobile eye
trackers, generally,

00:01:34.380 --> 00:01:36.300 align:middle line:84%
if the person moves
them a little bit

00:01:36.300 --> 00:01:41.130 align:middle line:84%
or if they're a bit sweaty, you
want to recalibrate sometimes.

00:01:41.130 --> 00:01:42.570 align:middle line:84%
So there's another
method that you

00:01:42.570 --> 00:01:45.810 align:middle line:84%
can use for calibrating
the mobile eye trackers.

00:01:45.810 --> 00:01:48.120 align:middle line:84%
If you don't have
a screen available

00:01:48.120 --> 00:01:51.780 align:middle line:84%
or you don't want to use a
screen, instead of presenting

00:01:51.780 --> 00:01:54.450 align:middle line:84%
calibration targets on
a screen or a projector,

00:01:54.450 --> 00:01:56.460 align:middle line:84%
you can choose
locations in the world

00:01:56.460 --> 00:01:58.980 align:middle line:84%
and ask the person
to focus their eyes

00:01:58.980 --> 00:02:04.110 align:middle line:84%
on several specific locations,
usually solid, fixed points.

00:02:04.110 --> 00:02:06.390 align:middle line:84%
And then use the
software that you're

00:02:06.390 --> 00:02:10.889 align:middle line:84%
using to capture the
eye tracking data

00:02:10.889 --> 00:02:13.350 align:middle line:84%
to indicate exactly where
the person is supposed

00:02:13.350 --> 00:02:15.090 align:middle line:90%
to be looking right now.

00:02:15.090 --> 00:02:19.350 align:middle line:84%
And this effectively carries
out the same procedure.

00:02:19.350 --> 00:02:20.070 align:middle line:90%
Yeah.

00:02:20.070 --> 00:02:23.340 align:middle line:84%
This was our brief explanation
about the calibration

00:02:23.340 --> 00:02:27.150 align:middle line:84%
of stationary eye trackers
and mobile eye trackers.

00:02:27.150 --> 00:02:28.560 align:middle line:84%
I hope you have
learned something

00:02:28.560 --> 00:02:31.860 align:middle line:84%
from this introduction
to eye tracking.

00:02:31.860 --> 00:02:34.280 align:middle line:90%
And good luck with your course.

00:02:34.280 --> 00:02:45.000 align:middle line:90%