WEBVTT

00:00.000 --> 00:05.000
Okay.

00:05.000 --> 00:09.000
Thank you.

00:09.000 --> 00:14.000
Well, I'll start with a little bit unplanned

00:14.000 --> 00:18.000
Pramble, but I just can't resist.

00:18.000 --> 00:20.000
Here it is.

00:20.000 --> 00:25.000
Approximately half a year ago, I found a really strange

00:25.000 --> 00:30.000
sheet of paper, which is seen on the photo.

00:30.000 --> 00:32.000
It's a typical shopping list.

00:32.000 --> 00:36.000
And some items are crossed out, like they were already bought.

00:36.000 --> 00:42.000
Funny moment, items, the keywords were put into language.

00:42.000 --> 00:48.000
Just in case if I forget one of the languages, probably I don't know.

00:48.000 --> 00:51.000
The list is written by my own hand.

00:51.000 --> 00:54.000
And some items are just unfamiliar.

00:54.000 --> 00:57.000
And some items were rarely strange.

00:57.000 --> 01:00.000
Frankly speaking, only the first item was familiar.

01:00.000 --> 01:05.000
The orange timberland boots, which was crossed out, like bought,

01:05.000 --> 01:07.000
but I never bought them.

01:07.000 --> 01:12.000
I don't know what is this about, just some visual feeling.

01:12.000 --> 01:16.000
And what about other items, just a few here,

01:17.000 --> 01:21.000
pink scarf for autumn.

01:21.000 --> 01:25.000
Maybe, well, sounds like a gift, but for whom.

01:25.000 --> 01:31.000
Some strange devices, Frank Hitler, I don't know what is this,

01:31.000 --> 01:37.000
or Christmas tree T-shirt, which is non-typical for winter, you know.

01:37.000 --> 01:43.000
And four flip flops with feathers.

01:43.000 --> 01:46.000
No comments.

01:46.000 --> 01:51.000
Drum for my nephew looks more or less okay, but I have no nephews.

01:51.000 --> 01:55.000
So this word mine is extremely strange.

01:55.000 --> 02:01.000
Then a rabbit shaped fond case will also be without comments.

02:01.000 --> 02:09.000
And the last and the most strong item is obviously protective helmet for a cat.

02:09.000 --> 02:17.000
No, I started even to file a pink ladies from my contact list,

02:17.000 --> 02:23.000
asking if they know anything about pink scarf for autumn.

02:23.000 --> 02:27.000
What surprised me, a little bit,

02:27.000 --> 02:32.000
two friends said they were not surprised by my shopping list,

02:32.000 --> 02:37.000
which is probably something to think about it later.

02:37.000 --> 02:44.000
Well, the end of the preamble, it took ten minutes of very unusual feelings.

02:44.000 --> 02:48.000
You know, maybe Dr. Jackl have felt something like this,

02:48.000 --> 02:51.000
when he first discovered Mr. Hyde.

02:51.000 --> 02:55.000
But fortunately, my Mr. Hyde was much more pleasant person,

02:55.000 --> 03:01.000
because you know, real Mr. Hyde would never wear four flip flops with feathers.

03:01.000 --> 03:02.000
And why did?

03:02.000 --> 03:06.000
Well, ten minutes later, I recalled what is this.

03:06.000 --> 03:13.000
It was a list of tasks for a small eye tracking research on online shopping patterns.

03:13.000 --> 03:18.000
And well, as you may guess, the magic was immediately gone,

03:18.000 --> 03:25.000
but it gave me a really strong motivation to give a talk on eye tracking.

03:25.000 --> 03:29.000
And that's actually the talk.

03:30.000 --> 03:34.000
How we can use eye tracking in user experience research.

03:34.000 --> 03:42.000
First, we can track the sources of UX problems like our elements are visible.

03:42.000 --> 03:46.000
What's the focal points of attention of the end user?

03:46.000 --> 03:52.000
Maybe discover some distractors, which attract a lot of gaze points.

03:52.000 --> 03:57.000
Then, visual search strategies, which we will a little bit speak about later,

03:57.000 --> 04:02.000
is also important thing, and as well as evidence-cunning patterns.

04:02.000 --> 04:07.000
And finally, it's really good if you want to make a bit testing of two solutions.

04:07.000 --> 04:12.000
Maybe the modified and original version of the interface or whatever.

04:12.000 --> 04:18.000
And speaking about the eye tracking technology by itself,

04:19.000 --> 04:24.000
typical eye tracker has several cameras, which are built into it,

04:24.000 --> 04:30.000
and infrared lamps, which are targeted into your eyes.

04:30.000 --> 04:37.000
So those cameras are recording video with your eyes reflecting the infrared light.

04:37.000 --> 04:45.000
It's like red eye effect on photos, but in infrared range, so you notice nothing.

04:45.000 --> 04:52.000
And then they make some geometric calculations to figure out where you are staring at on the screen.

04:52.000 --> 05:01.000
Sounds rather easy, and by the way, normally, when you make eye tracking,

05:01.000 --> 05:09.000
what matters are fixations when you look at something and sockets, which means jumping from one gaze point to another.

05:10.000 --> 05:13.000
You see the text-cunning pattern here.

05:13.000 --> 05:18.000
Also, it would be really good to take into account that you definitely,

05:18.000 --> 05:27.000
well, your gaze is not arrow, which makes a tiny pinhole.

05:27.000 --> 05:32.000
You see a little bit more than a defecation point shows,

05:32.000 --> 05:36.000
just through the resolution quickly decreases around.

05:36.000 --> 05:47.000
So we still draw a fixations, but this graph of fixations doesn't mean that the user sees nothing aside of them.

05:47.000 --> 05:49.000
It's just the point of focus.

05:49.000 --> 05:53.000
And extreme cases of eye tracking, the very first one,

05:53.000 --> 06:00.000
a professional eye tracker, sometimes, had mounted nowadays stationery once again popular.

06:00.000 --> 06:02.000
They are really expensive.

06:03.000 --> 06:06.000
Well, some of them are really usable on the Linux,

06:06.000 --> 06:12.000
because they simulate network server with some JSON data or something like this,

06:12.000 --> 06:15.000
maybe XML or anything.

06:15.000 --> 06:20.000
But on the opposite, well, the price is not very friendly.

06:20.000 --> 06:21.000
Absolutely not.

06:21.000 --> 06:26.000
On the opposite side, you get webcames based software eye trackers,

06:26.000 --> 06:29.000
because, well, you know, your laptop has web camera,

06:29.000 --> 06:34.000
so you can use open civil library to track or something similar to track your eyes as well.

06:34.000 --> 06:43.000
The main problem is that the resolution of the eye tracking will be to times lower than the resolution of your camera.

06:43.000 --> 06:49.000
And you know, focus, screen, and hopefully an HD camera,

06:49.000 --> 06:54.000
and HD divided by two, you know, the problem.

06:55.000 --> 06:58.000
And a firm rate also makes sense.

06:58.000 --> 07:01.000
And always the middle way, the consumer eye trackers,

07:01.000 --> 07:04.000
gaming industry helps us a lot,

07:04.000 --> 07:08.000
because, well, they use eye trackers to budget some missiles,

07:08.000 --> 07:11.000
in computer games or anything like this,

07:11.000 --> 07:16.000
and you need enough precision and enough accuracy to target anything,

07:16.000 --> 07:21.000
to get advantage over gamers, which have no such device.

07:21.000 --> 07:25.000
So these devices have to be good enough.

07:25.000 --> 07:28.000
Otherwise gamers will not buy them.

07:28.000 --> 07:36.000
And sometimes, at least to be, company produces software development kits for Linux,

07:36.000 --> 07:39.000
and quickly abundance them.

07:39.000 --> 07:45.000
But, in a short period of time, when the HDK is available,

07:45.000 --> 07:48.000
it's more or less usable.

07:48.000 --> 07:51.000
And a little bit more about accuracy and precision.

07:51.000 --> 07:58.000
Accuracy means how different are actual and recorded eye movements,

07:58.000 --> 08:00.000
movements, not case points.

08:00.000 --> 08:06.000
And precision, how reliably the same gas point is reproduced.

08:06.000 --> 08:09.000
So accuracy and precision is something to take into account,

08:09.000 --> 08:13.000
when you choose eye tracking to use,

08:13.000 --> 08:16.000
and speaking about frame rate on the screen,

08:16.000 --> 08:21.000
you see two bottom items, professional eye trackers from toy

08:21.000 --> 08:24.000
and three top items, gaming eye trackers.

08:24.000 --> 08:28.000
And you see the frame rate is really good.

08:28.000 --> 08:31.000
Otherwise, they will not work.

08:31.000 --> 08:33.000
So, how to use consumer eye tracker.

08:33.000 --> 08:40.000
First, it's probably running with outdated Linux control,

08:41.000 --> 08:44.000
because of old SDK.

08:44.000 --> 08:48.000
So, there is highly probable that you will use two laptops.

08:48.000 --> 08:52.000
You know, you put eye tracker on one laptop, computer one,

08:52.000 --> 08:55.000
and connected to another laptop computer two

08:55.000 --> 08:59.000
with some outdated Linux control and SDK.

08:59.000 --> 09:01.000
If the screens are different,

09:01.000 --> 09:05.000
you will probably use some desktop software

09:05.000 --> 09:09.000
to show calibration screen on the main computer.

09:09.000 --> 09:12.000
The one you choose for testing,

09:12.000 --> 09:15.000
and after that, the second computer just collects data,

09:15.000 --> 09:22.000
and works like data collection station.

09:22.000 --> 09:27.000
If speaking about the software on the eye tracker,

09:27.000 --> 09:30.000
you may think that, well,

09:30.000 --> 09:34.000
webcams of laptops are not very good,

09:34.000 --> 09:38.000
but what about the smartphones,

09:38.000 --> 09:41.000
which definitely are really strong in resolution,

09:41.000 --> 09:44.000
and have from time to time a really good frame rate?

09:44.000 --> 09:46.000
Yes, that's possible.

09:46.000 --> 09:48.000
You can use smartphone in the same way.

09:48.000 --> 09:52.000
You can use some Android app to stream

09:52.000 --> 09:57.000
with smartphone scanner as a webcam over the internet,

09:57.000 --> 10:00.000
or you can use some features of Android.

10:00.000 --> 10:05.000
It's little bit more tricky to pretend it to be a USB camera.

10:06.000 --> 10:09.000
Anyway, that's probably the better approach

10:09.000 --> 10:14.000
taken into account the typical webcams of laptop,

10:14.000 --> 10:17.000
and about the collected data.

10:17.000 --> 10:20.000
The optimistic version looks like these.

10:20.000 --> 10:22.000
So number of fixations,

10:22.000 --> 10:25.000
duration of each fixation,

10:25.000 --> 10:28.000
viewing speed, even pupil diameter,

10:28.000 --> 10:30.000
and a lot of other things,

10:30.000 --> 10:34.000
which are metrics of a kilogram,

10:34.000 --> 10:38.000
but may be not too easy to collect,

10:38.000 --> 10:40.000
with consumer grade attackers.

10:40.000 --> 10:45.000
Fortunately, UX research is not something related to medicine.

10:45.000 --> 10:48.000
We need not too much, by the way.

10:48.000 --> 10:51.000
And visualization, by the way,

10:51.000 --> 10:53.000
can be done in two ways.

10:53.000 --> 10:55.000
The gaze plot,

10:55.000 --> 10:58.000
the graph of fixation.

10:58.000 --> 11:02.000
It's sometimes uses size of the node,

11:02.000 --> 11:05.000
and numbering, and other additional things,

11:05.000 --> 11:07.000
and hit map,

11:07.000 --> 11:10.000
which shows the color temperatures

11:10.000 --> 11:12.000
to show distribution of fixations

11:12.000 --> 11:14.000
when you have too much of them.

11:14.000 --> 11:16.000
So the short-chain data

11:16.000 --> 11:19.000
are rather good for gaze plots.

11:19.000 --> 11:21.000
Here is an example.

11:21.000 --> 11:22.000
A sample app,

11:22.000 --> 11:24.000
with two windows on the right one,

11:24.000 --> 11:26.000
you have a geometric figure,

11:26.000 --> 11:28.000
and you should find it in the left window,

11:28.000 --> 11:31.000
among 25 geometric figures.

11:31.000 --> 11:36.000
And you see where a short gaze plot,

11:36.000 --> 11:39.000
and built with graph is,

11:39.000 --> 11:42.000
and on the right, you see actual graph is file,

11:42.000 --> 11:44.000
which actually builds the graph.

11:44.000 --> 11:47.000
It's rather simple,

11:47.000 --> 11:52.000
but if you have longer experiment,

11:53.000 --> 11:56.000
gaze plot becomes not very usable.

11:56.000 --> 11:59.000
For example, here,

11:59.000 --> 12:02.000
we have built a simple gaze plot with

12:02.000 --> 12:04.000
an octave,

12:04.000 --> 12:06.000
three lines, by the way.

12:06.000 --> 12:08.000
Taking into account

12:08.000 --> 12:11.000
that you get coordinates of fixation

12:11.000 --> 12:14.000
from the CSV file, easy.

12:14.000 --> 12:16.000
Just,

12:16.000 --> 12:19.000
gaze plot is not very good here,

12:19.000 --> 12:21.000
by the way.

12:21.000 --> 12:23.000
So hitmaps,

12:23.000 --> 12:24.000
even,

12:24.000 --> 12:25.000
well,

12:25.000 --> 12:27.000
not three lines,

12:27.000 --> 12:28.000
ten lines,

12:28.000 --> 12:29.000
of course,

12:29.000 --> 12:30.000
in octave,

12:30.000 --> 12:32.000
will bring you these,

12:32.000 --> 12:33.000
with additional function,

12:33.000 --> 12:36.000
which actually does the magic.

12:36.000 --> 12:38.000
And,

12:38.000 --> 12:40.000
say,

12:40.000 --> 12:42.000
about hitmaps,

12:42.000 --> 12:46.000
here you see few examples of the same test,

12:47.000 --> 12:50.000
with geometric figures,

12:50.000 --> 12:52.000
detected.

12:52.000 --> 12:53.000
And,

12:53.000 --> 12:57.000
here you clearly see the scanning patterns.

12:57.000 --> 12:59.000
And this approach,

12:59.000 --> 13:00.000
by the way,

13:00.000 --> 13:02.000
is really good when you need to know

13:02.000 --> 13:04.000
if the control panel is,

13:04.000 --> 13:06.000
scanned element by element,

13:06.000 --> 13:08.000
because of the miller wallet,

13:08.000 --> 13:09.000
a rule,

13:09.000 --> 13:11.000
and bed grouping,

13:11.000 --> 13:13.000
or whether the person detects

13:13.000 --> 13:14.000
group of items,

13:14.000 --> 13:17.000
and then detects the item of the group.

13:19.000 --> 13:22.000
About interpreting the data,

13:22.000 --> 13:24.000
definitely you will take into account

13:24.000 --> 13:25.000
zones of the screen,

13:25.000 --> 13:28.000
because there will be destruction zones,

13:28.000 --> 13:30.000
something not related to the task,

13:30.000 --> 13:33.000
and maybe the working area,

13:33.000 --> 13:35.000
if that's some software,

13:35.000 --> 13:36.000
with toolbar,

13:36.000 --> 13:38.000
then there will be the toolbar area,

13:38.000 --> 13:42.000
and you can figure out how it works.

13:42.000 --> 13:43.000
By the way,

13:43.000 --> 13:44.000
here,

13:44.000 --> 13:46.000
it's important to take into account

13:46.000 --> 13:48.000
that software I triggers,

13:48.000 --> 13:51.000
which we will speak a little bit more in a moment,

13:51.000 --> 13:54.000
and hardware I triggers may have

13:54.000 --> 13:56.000
different local systems of coordinates,

13:56.000 --> 13:57.000
because,

13:57.000 --> 13:58.000
well,

13:58.000 --> 13:59.000
the origin of top-backed triggers

13:59.000 --> 14:03.000
is not in the top-left corner,

14:03.000 --> 14:06.000
but in the top-left bottom corner.

14:06.000 --> 14:07.000
So,

14:07.000 --> 14:11.000
maybe you will have to turn around the picture.

14:12.000 --> 14:14.000
But, for example,

14:14.000 --> 14:15.000
here,

14:15.000 --> 14:16.000
you can see,

14:16.000 --> 14:19.000
example of two office use.

14:19.000 --> 14:22.000
I reduce the presentation a little bit,

14:22.000 --> 14:25.000
so we don't have actual screen shots,

14:25.000 --> 14:30.000
but that's checking the scanning patterns

14:30.000 --> 14:32.000
on the ribbon-like interface,

14:32.000 --> 14:35.000
introduced several years ago,

14:35.000 --> 14:38.000
in liberal office following the Microsoft office use.

14:38.000 --> 14:39.000
And,

14:39.000 --> 14:41.000
one more important thing is,

14:41.000 --> 14:44.000
about comparing the hardware and

14:44.000 --> 14:46.000
canvas I triggers.

14:46.000 --> 14:47.000
Here,

14:47.000 --> 14:49.000
you see a good example,

14:49.000 --> 14:51.000
which shows you how good

14:51.000 --> 14:53.000
is webcam-based I trigger,

14:53.000 --> 14:58.000
comparison to the mid-range gaming one.

14:58.000 --> 15:02.000
Those points are this plot

15:02.000 --> 15:05.000
or just fixations graph,

15:05.000 --> 15:08.000
obtained from the software I trigger,

15:08.000 --> 15:10.000
the webgays of Gs,

15:10.000 --> 15:14.000
which I've recommended several slides ago.

15:14.000 --> 15:16.000
And, on the right,

15:16.000 --> 15:20.000
you see the gaze plot taken by a staff.

15:20.000 --> 15:22.000
You can notice,

15:22.000 --> 15:23.000
sometimes,

15:23.000 --> 15:25.000
the figures are more or less,

15:25.000 --> 15:27.000
similarly looking,

15:27.000 --> 15:28.000
but,

15:28.000 --> 15:29.000
definitely,

15:29.000 --> 15:33.000
the accuracy and precision are not quite in good.

15:33.000 --> 15:34.000
And,

15:34.000 --> 15:35.000
what's more important,

15:35.000 --> 15:37.000
missing points.

15:37.000 --> 15:38.000
You know,

15:38.000 --> 15:40.000
you need a higher frame rate than

15:40.000 --> 15:43.000
25 frames per second,

15:43.000 --> 15:45.000
to get all fixations.

15:45.000 --> 15:46.000
And, here,

15:46.000 --> 15:48.000
we used a rather good resolution,

15:48.000 --> 15:50.000
but,

15:50.000 --> 15:52.000
the example was a little bit spoiled

15:52.000 --> 15:55.000
by a normal frame rate.

15:55.000 --> 15:58.000
So, you may be more likely

15:58.000 --> 16:02.000
with a good smartphone camera,

16:02.000 --> 16:04.000
not in terms of resolution,

16:04.000 --> 16:07.000
but the frame rate is actually what works.

16:07.000 --> 16:09.000
And, a bonus thing,

16:09.000 --> 16:12.000
what about mouse hit maps.

16:12.000 --> 16:13.000
Well,

16:13.000 --> 16:14.000
technically,

16:14.000 --> 16:16.000
building hit map from CSV is

16:16.000 --> 16:19.000
almost similar task as

16:19.000 --> 16:24.000
building hit map from CSV,

16:24.000 --> 16:26.000
which I mean,

16:26.000 --> 16:29.000
CSV taken from I trigger,

16:29.000 --> 16:31.000
or CSV taken from the course of our positions.

16:32.000 --> 16:34.000
By the way,

16:34.000 --> 16:38.000
only one line command does

16:38.000 --> 16:42.000
these recording of mouse positions in the Linux.

16:42.000 --> 16:43.000
Sorry,

16:43.000 --> 16:45.000
it's long command.

16:45.000 --> 16:46.000
And,

16:46.000 --> 16:48.000
I'm pretending it's a one-liner,

16:48.000 --> 16:49.000
but,

16:49.000 --> 16:51.000
not very small one-liner,

16:51.000 --> 16:52.000
but still,

16:52.000 --> 16:55.000
the comparison.

16:55.000 --> 16:57.000
On the right,

16:57.000 --> 16:59.000
you see some documentation.

16:59.000 --> 17:00.000
Well,

17:00.000 --> 17:01.000
yeah,

17:01.000 --> 17:03.000
it was usability research for

17:03.000 --> 17:05.000
particular documentation.

17:05.000 --> 17:06.000
And here,

17:06.000 --> 17:08.000
we were tracking whether

17:08.000 --> 17:10.000
users are using

17:10.000 --> 17:12.000
actually menus to find things

17:12.000 --> 17:15.000
or they used the search field on the top,

17:15.000 --> 17:17.000
and you see how gaze maps

17:17.000 --> 17:20.000
are different from mouse maps.

17:20.000 --> 17:22.000
Well,

17:22.000 --> 17:25.000
software I tracking is definitely better.

17:26.000 --> 17:28.000
Funny moment,

17:28.000 --> 17:31.000
some people are selling these,

17:31.000 --> 17:32.000
you know,

17:32.000 --> 17:37.000
service for making mouse maps for your plot,

17:37.000 --> 17:40.000
saying that hit maps will help you.

17:40.000 --> 17:41.000
Well,

17:41.000 --> 17:42.000
in this task,

17:42.000 --> 17:44.000
at least we can track whether

17:44.000 --> 17:47.000
a user navigates where menu or not.

17:47.000 --> 17:50.000
It's not bad as additional,

17:50.000 --> 17:51.000
you know,

17:51.000 --> 17:52.000
actually,

17:52.000 --> 17:53.000
two,

17:53.000 --> 17:55.000
and it's easy to gather,

17:55.000 --> 17:56.000
but,

17:56.000 --> 17:58.000
obviously not a substitution.

17:58.000 --> 17:59.000
And,

17:59.000 --> 18:00.000
so,

18:00.000 --> 18:02.000
what about conclusions?

18:02.000 --> 18:04.000
Well,

18:04.000 --> 18:10.000
despite the typical opinion,

18:10.000 --> 18:13.000
I tracking in Linux is not very difficult,

18:13.000 --> 18:14.000
I would say,

18:14.000 --> 18:17.000
and the user will create

18:18.000 --> 18:23.000
a good performance for usability tasks,

18:23.000 --> 18:25.000
for usability research tasks,

18:25.000 --> 18:26.000
maybe not for medicine,

18:26.000 --> 18:27.000
but fortunately,

18:27.000 --> 18:30.000
we are in much easier situation.

18:30.000 --> 18:32.000
Web comes,

18:32.000 --> 18:33.000
okay,

18:33.000 --> 18:34.000
more or less,

18:34.000 --> 18:36.000
if you have really good camera,

18:36.000 --> 18:38.000
with good frame rate and resolution,

18:38.000 --> 18:39.000
and so on,

18:39.000 --> 18:40.000
but,

18:40.000 --> 18:41.000
fortunately,

18:41.000 --> 18:43.000
the easiest part is data visualization.

18:43.000 --> 18:46.000
Well,

18:47.000 --> 18:48.000
and,

18:48.000 --> 18:49.000
you know,

18:49.000 --> 18:51.000
everything about capturing a mouse

18:51.000 --> 18:52.000
from the previous slide,

18:52.000 --> 18:53.000
and probably that's

18:53.000 --> 18:55.000
the cat in protective helmet,

18:55.000 --> 18:56.000
I guess so.

18:56.000 --> 18:58.000
I'm not pretty sure,

18:58.000 --> 19:00.000
which protective helmet for cat

19:00.000 --> 19:01.000
was bought,

19:01.000 --> 19:03.000
virtually bought in this research.

19:03.000 --> 19:04.000
So,

19:04.000 --> 19:05.000
that's,

19:05.000 --> 19:07.000
the last slide,

19:07.000 --> 19:10.000
just not to finish with the cat photo.

19:10.000 --> 19:13.000
Probably that's a happy user,

19:14.000 --> 19:18.000
which have just finished usability research,

19:18.000 --> 19:21.000
and that's it.

19:21.000 --> 19:23.000
Thank you for.

19:23.000 --> 19:32.000
Thank you for.

19:32.000 --> 19:33.000
Yeah, a question.

19:33.000 --> 19:34.000
Yeah.

19:34.000 --> 19:35.000
So,

19:35.000 --> 19:37.000
attract my users,

19:37.000 --> 19:38.000
I gaze.

19:38.000 --> 19:39.000
I've got a list map,

19:39.000 --> 19:40.000
I've got a nice eat map,

19:41.000 --> 19:42.000
some gays,

19:42.000 --> 19:43.000
diagrams.

19:43.000 --> 19:44.000
What do I do with it?

19:44.000 --> 19:45.000
Oh.

19:45.000 --> 19:48.000
Good question.

19:48.000 --> 19:49.000
And,

19:49.000 --> 19:52.000
what to do with the diagrams,

19:52.000 --> 19:56.000
when you collect them?

19:56.000 --> 19:57.000
Yeah.

19:57.000 --> 19:59.000
That's a typical question,

19:59.000 --> 20:00.000
for example,

20:00.000 --> 20:02.000
if you have built personas,

20:02.000 --> 20:03.000
for your resource,

20:03.000 --> 20:05.000
what to do with those personas.

20:05.000 --> 20:08.000
I have a slide,

20:08.000 --> 20:10.000
which says how,

20:10.000 --> 20:12.000
those diagrams,

20:12.000 --> 20:15.000
allow you to see what's working or what's not.

20:15.000 --> 20:17.000
For example,

20:17.000 --> 20:18.000
well,

20:18.000 --> 20:19.000
you know,

20:19.000 --> 20:20.000
charismatic example of

20:20.000 --> 20:24.000
whether users are seeing your bananas or not.

20:24.000 --> 20:25.000
Or,

20:25.000 --> 20:29.000
whether they use scanning or just,

20:29.000 --> 20:30.000
well,

20:30.000 --> 20:32.000
for example,

20:32.000 --> 20:34.000
in KDE version 3,

20:34.000 --> 20:36.000
they try to use not-spaces

20:37.000 --> 20:39.000
to separate groups of icons on two bars,

20:39.000 --> 20:41.000
but color differentiation,

20:41.000 --> 20:42.000
which didn't work.

20:42.000 --> 20:43.000
And,

20:43.000 --> 20:45.000
if they use eye tracking research,

20:45.000 --> 20:48.000
they would quickly see that users are scanning

20:48.000 --> 20:50.000
the two bars of conqueror,

20:50.000 --> 20:52.000
the Trinity desktop nowadays.

20:52.000 --> 20:55.000
It's actually the fork of KDE 3.

20:55.000 --> 20:56.000
So,

20:56.000 --> 20:58.000
it's still there.

20:58.000 --> 20:59.000
And,

20:59.000 --> 21:00.000
well,

21:00.000 --> 21:02.000
aside of scanning patterns,

21:02.000 --> 21:03.000
definitely,

21:03.000 --> 21:07.000
you will do another tracking errors,

21:07.000 --> 21:08.000
time tracking,

21:08.000 --> 21:09.000
obviously,

21:09.000 --> 21:11.000
may be some questionaries,

21:11.000 --> 21:14.000
but it's really helpful to see

21:14.000 --> 21:16.000
what users see,

21:16.000 --> 21:21.000
and how users see your resource.

21:21.000 --> 21:23.000
So, that's the answer.

21:23.000 --> 21:24.000
And,

21:24.000 --> 21:25.000
yeah,

21:25.000 --> 21:26.000
the slide.

21:26.000 --> 21:28.000
Yeah, please.

21:28.000 --> 21:29.000
You know,

21:29.000 --> 21:31.000
a tiny project that uses the

21:31.000 --> 21:32.000
eye icon,

21:32.000 --> 21:35.000
or that's a lot of search for a piece of news.

21:35.000 --> 21:37.000
I don't know a project,

21:37.000 --> 21:40.000
which uses this camera for eye tracking,

21:40.000 --> 21:41.000
but fortunately,

21:41.000 --> 21:42.000
it's not needed.

21:42.000 --> 21:44.000
If you use webgas or gs,

21:44.000 --> 21:45.000
this one,

21:45.000 --> 21:51.000
which was present in this comparison,

21:51.000 --> 21:53.000
and the one,

21:53.000 --> 21:56.000
which we have,

21:56.000 --> 21:58.000
URL here.

21:58.000 --> 22:00.000
Well,

22:00.000 --> 22:02.000
there are several projects,

22:02.000 --> 22:04.000
which uses camera,

22:04.000 --> 22:05.000
USB camera,

22:05.000 --> 22:06.000
or web camera,

22:06.000 --> 22:07.000
or laptop,

22:07.000 --> 22:08.000
or any other camera,

22:08.000 --> 22:09.000
as,

22:09.000 --> 22:11.000
I track it,

22:11.000 --> 22:12.000
this is one,

22:12.000 --> 22:14.000
which is still alive,

22:14.000 --> 22:16.000
more than others.

22:16.000 --> 22:17.000
And,

22:17.000 --> 22:19.000
it's absolutely universal,

22:19.000 --> 22:21.000
because it's JavaScript.

22:21.000 --> 22:23.000
It brings,

22:23.000 --> 22:25.000
to a browser console,

22:25.000 --> 22:27.000
all fixations,

22:27.000 --> 22:28.000
it records.

22:29.000 --> 22:32.000
If you are working not with the web browser,

22:32.000 --> 22:33.000
a product,

22:33.000 --> 22:34.000
a lot with the website,

22:34.000 --> 22:35.000
for example,

22:35.000 --> 22:36.000
you just use full screen,

22:36.000 --> 22:37.000
the browser,

22:37.000 --> 22:40.000
as a substrate for your screen.

22:40.000 --> 22:41.000
And,

22:41.000 --> 22:44.000
it will still record all fixations on your screen,

22:44.000 --> 22:46.000
even if you do not see it.

22:46.000 --> 22:49.000
You know,

22:49.000 --> 22:53.000
this nice cameras are just not needed to solve,

22:53.000 --> 22:54.000
I mean,

22:54.000 --> 22:56.000
this sophisticated cameras with face ID,

22:56.000 --> 22:57.000
and so on,

22:57.000 --> 22:59.000
not needed to solve this problem.

22:59.000 --> 23:00.000
You see,

23:00.000 --> 23:01.000
this one,

23:01.000 --> 23:05.000
really vintage solution based on OpenCV,

23:05.000 --> 23:09.000
worked quite well

23:09.000 --> 23:11.000
without additional technologies.

23:11.000 --> 23:14.000
Maybe,

23:14.000 --> 23:16.000
if you would like

23:16.000 --> 23:19.000
to track eyes

23:19.000 --> 23:22.000
in absolutely dark environment,

23:22.000 --> 23:24.000
then those cameras will be helpful,

23:25.000 --> 23:27.000
but I probably think you will not.

23:27.000 --> 23:28.000
So,

23:28.000 --> 23:30.000
this sophisticated hardware is just not needed.

23:30.000 --> 23:32.000
What needed is frame rate,

23:32.000 --> 23:33.000
and resolution,

23:33.000 --> 23:35.000
and those built-in cameras,

23:35.000 --> 23:37.000
typically have not very high resolution,

23:37.000 --> 23:39.000
and not very high frame rate.

23:42.000 --> 23:43.000
Oh,

23:43.000 --> 23:44.000
one more moment.

23:44.000 --> 23:46.000
If you use hardware,

23:46.000 --> 23:48.000
I track or take into account that,

23:48.000 --> 23:49.000
for gaming,

23:49.000 --> 23:51.000
you have no

23:52.000 --> 23:55.000
possibility to legally store

23:55.000 --> 23:58.000
coordinates or fixations to a file,

23:58.000 --> 24:02.000
because they are supposed to be personal data.

24:02.000 --> 24:03.000
Frankly speaking,

24:03.000 --> 24:06.000
because they will still be self-operative license for this.

24:06.000 --> 24:07.000
But,

24:07.000 --> 24:08.000
they have an example,

24:08.000 --> 24:11.000
which is able to show coordinates on the screen,

24:11.000 --> 24:14.000
and just using

24:14.000 --> 24:15.000
input-out,

24:15.000 --> 24:18.000
with the direction from this print to console,

24:18.000 --> 24:19.000
example,

24:19.000 --> 24:21.000
you can build

24:21.000 --> 24:24.000
fog map or heat map

24:24.000 --> 24:27.000
on the go without any problems.

24:30.000 --> 24:31.000
Thank you very much.

24:31.000 --> 24:32.000
Thank you once again.

24:32.000 --> 24:33.000
Thank you.