[tt] [croquet-user] Eyetracking VS BCI (was: Impressive, soon-to-come brain control headset)

Thu Feb 21 10:31:53 UTC 2008

----- Forwarded message from Aaron Brancotti <aaron.brancotti at srlabs.it> -----

From: Aaron Brancotti <aaron.brancotti at srlabs.it>
Date: Thu, 21 Feb 2008 10:55:54 +0100
To: croquet-user at duke.edu
Subject: [croquet-user] Eyetracking VS BCI (was:  Impressive, soon-to-come brain control
 headset)
User-Agent: Thunderbird 2.0.0.9 (Windows/20071031)
Reply-To: croquet-user at duke.edu,
	Aaron Brancotti <aaron.brancotti at srlabs.it>

Hi all from Italy and from a list newby (but a long-time addicted VR 
developer, I was working with W-Industries Virtuality stuff back in 1992),

just my 2 cents:

sure using directly electrical signals from our nervous system is 
fascinating, but I think it is still too cumbersome. And the EPOC and 
BCI (Brain Computer Interface) stuff, YES, you can do that, and it 
somewhat works, but not right out of the box. You need HW, and you must 
wear it, and you need training, and...
>A more down to earth look at input devices is being done by a friend
>of mine called Sam [snip] and contains gyros, accellerometers, a squidgy
>pressure sensor thingy and a microphone.
>  
your friend's HW work is great, but it will not be a question of just 
hardware alone. Sure, some low-cost , feature-packed stuff like that (or 
the wiimote) can help in making a computer more USABLE (keep in mind 
this word! :) ), but IMHO the real shift will come from non-intrusive, 
"natural" technologies like eyetracking and voice-recognition (as your 
friend states at the bottom of the page). And in this, I see Fix8 much 
more interesting than other do, it seems:
>>The way that I think of fix8 (and Logitech orbitcam) as being
>> comparable to emotiv is that they all make use of a lot of user-input
>> information that is otherwise being thrown away.  The camera stuff
>> doesn't happen to be doing anything semantic with those gestures
>> right now, but they are capturing stuff.
>>    
well, you are probably right, meaning that Fix8 probably just does some 
feature recognition (eyes position, nose position etc) and translates 
1:1 the movements of those features on a 3D model, but applying 
semantics is quite near to that...

>> > I think fix8 appears to be solving a different problem.  [snip]  Mice 
>> and keyboards are
>> > not very usable as inputs for wearable computers, and the disabled
>> > have trouble with mice and keyboards as well.
>>    
some kind of disabilities, like lateral amiotrophic sclerosis, would 
pose problems even for electrodes-based stuff like the EPOC: such kind 
of disabilities completely blocks your body, and even wearing something 
on your head would be cumbersome (you must always rest your head), 
without mentioning that it would require specialized knowledge and 
assistance and IMHO could be "scary" for a lot of people. On the other 
hand, technologies like eyetracking are completely unobtrusive and can 
be used effectively (with a COMPLETE REDESIGN of the user interface, 
obviously) to control a computer with just your eyes. The company for 
which I work develops and sell a whole SW suite comprising a mail 
client, a web browser (mozilla-based), some specialized on-screen 
keyboards, a Skype client, vocal synthesis, an e-book reader, a small 
"write" app, a simplified "file system" and some more stuff. Problem now 
is the HW (the eyetracker), which is VERY expensive, because the makers 
are market leaders. But the tecnology itself is no secret, it is just 
they were first and are still the best and have the best algorithms 
since now. In different declinations, using eyetracking+voice 
recognition to achieve multimodal interaction is ALREADY an impressive 
application, and IMHO is absolutely superior to neurological/biofeedback 
solutions right now. I hardly figure someone able to write an email and 
surf the Web with just a BCI device and ten minutes setup and training.

In this scenario, Fix8 will NOT even bore the professional eyetracking 
hardware market with a 300$ SW working on a webcam, but they are 
developing knowledge in a very promising direction. After all, 
eyetracking is a matter of models, and filters, and fast image 
recognition, and maths... MAYBE a day they will develop some low-cost 
stuff able to do something similar to what you now have to pay 20K$ to 
get done. We are used to see software making miracles, are we not?

Well, on the other hand, the last frontier of neurological approach is 
that Matrix jack behind your neck, sure... but I will NOT beta-test it. :)

Aaron Brancotti aka Babele Dunnit

PS: I re-descovered the Croquet environment after some years, I read the 
FAQ and Overview and I immediately joined the list... men, you are doing 
it right. Hats off.

My two cents also about virtual worlds (so, this sums up to FOUR cents.. 
you getting rich, huh?)

http://www.nextverso.org/?p=6

Me and my friend wrote this a week BEFORE I re-descovered Croquet and I 
saw how many steps forward you took. I will definitely post something on 
NextVerso about Croquet, and why Croquet is the right way to go to clear 
our cyberspaces from that messy SecondLife stuff and build something 
really good.

----- End forwarded message -----
-- 
Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org
______________________________________________________________
ICBM: 48.07100, 11.36820 http://www.ativel.com http://postbiota.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE