Tobii Releases Eye-Controlled Mouse For PCs 67
Zothecula writes "Sweden's eye tracking and control innovator Tobii has announced the release of a stand-alone eye control device called PCEye. Like the laptop-based prototype system demonstrated at CeBIT last month, the device tracks eye movement and translates it into mouse cursor action onscreen. Positioned in front of a PC monitor and connected via USB, the company says that the technology is compatible with a wide range of software and has been primarily designed to improve computer interactivity for users with impaired motor skills, such as stroke victims." The estimated price range is around $7k, so yeah. Start saving your pennies for something cooler.
Mouse vs Cursor (Score:1)
Mouse is a piece of hardware. Cursor is the pointing device on the screen.
Re:Mouse vs Cursor (Score:2)
But... (Score:4, Funny)
But what if I don't want to just click on pictures of breasts?
Re:But... (Score:3)
Re:But... (Score:1)
Re:But... (Score:5, Funny)
Well, you could always try making eye contact occasionally. ;-)
Re:But... (Score:2)
Re:But... (Score:2)
You expect a slashdotter to make eye contact? Yhay's just wrong on so many levels!
Re:But... (Score:0)
Yes, but think in positive. At least your hands are free while you look at the nipples :D
Re:But... (Score:0)
Yes, this is a very bad technology for teleconferencing.
$7k?! (Score:3)
I wonder if someone is able to create something similar using an $20 webcam and some coding? :)
Re:$7k?! (Score:2)
i assure you it's the software thats the expensive part..
Re:$7k?! (Score:2)
Re:$7k?! (Score:1)
You're either joking, or have never worked in software development.
If it took several years and a whole lot of manpower to develop it, they need to recoup their past costs and pay for on-going maintenance and development, and pay their investors back. Oh, and of course pay for offices and support staff. It's not like the cost and time of development is negligible and then you just move onto selling it cheaply.
Umm, no. Until they recover costs they've already spent (plus other on-going costs), if they lower the price by several orders of magnitude, they never actually get to the point where they're being "profitable" in any meaningful sense of the word. If they do that up-front, they go broke.
Methinks you don't actually know WTF you're talking about. You should probably learn a little about how businesses operate, because pretty much everything you said is false.
Re:$7k?! (Score:1)
Which is all great and dandy until somebody comes along and releases the open source equivalent that they've been working on as their hobby.
Re:$7k?! (Score:2)
Which is why they hopefully have patents on what they've been working on.
Re:$7k?! (Score:2)
What, like this?
http://www.eyewriter.org/ [eyewriter.org]
Re:$7k?! (Score:2)
no - the software costs millions to manufacture.. and costs just pennies to duplicate.
the making it up in volume works but it depends on the market they are aiming for.. while i bet this does it's job and allows a stroke victim to use a computer i bet it isn't of the quality most people would want for gaming.. there for they stick to their market to make back their money.
Re:$7k?! (Score:2)
Re:$7k?! (Score:2)
you are correct but the point still stands - they need to recoup their cost - i'm sure if they thought it was good enough for mass consumption that they would have ventured down that route.
Re:$7k?! (Score:2)
Re:$7k?! (Score:2)
Re:$7k?! (Score:3, Informative)
I wonder if someone is able to create something similar using an $20 webcam and some coding? :)
It can be and I have done it. Its based on head movement but can be easily done with the eyes and a little bit coding revamped. Free VS 7grand. Seems fair to me.
http://reallifecod.webs.com/
Re:$7k?! (Score:1)
Re:$7k?! (Score:1)
S. Xavier-de-Souza, M. Van Dyck, J. Suykens and J. Vandewalle, "Fast and Robust Face Tracking for CNN Chips: Application to Wheelchair Driving", Int’l Workshop on Cellular Neural Networks and Their Applications, 2006.
Not exactly the same, but I think that is close. I have seen it working. Not $20 but way cheaper. Some video demos at ftp://wgs.esat.kuleuven.ac.be/sista/sdesouza/Thesis/
Re:$7k?! (Score:0)
Not quite. The Tobii system is a Pukinji image eyetracker/gazetracker. This requires capturing high resolution images of the eyes. Off axis IR illuminators are required to provide the glint. A $20 webcam will not have nearly the resolution or sensitivity needed for this application. Most commercial systems either roll their own camera or use a $1K+ PointGrey camera.
The gaze trackers referenced elsewhere herein are headtrackers not gazetrackers. They track the orientation of the users head, not where their eyes are fixated. So if you face forward but look to your left, the system will consider you looking forward, not left. More here:
http://en.wikipedia.org/wiki/Eye_tracking [wikipedia.org]
This company makes similar gazetracker/eyetracker to the Tobeii unit:
http://www.eyetechds.com/ [eyetechds.com]
Re:$7k?! (Score:1)
Re:$7k?! (Score:2)
I wonder if someone is able to create something similar using an $20 webcam and some coding? :)
I imagine that one advantage is the outsourcing of the image recognition to an external processor. Rather than using _your_ CPU for all of that math, you can use an external box for that. By using its own camera and its own circuitry, the only processing it needs to do on your computer is the mouse movement. (which hardly requires any power). Essentially, you can add the feature without slowing your computer down at all. (just speculation, though)
Also, I agree with the others. It's not expensive because of the hardware - it's expensive because they have to pay off their enormous R&D investment. (and because people with disabilities will pay it - I would guess that mouse movement with your eyes is much better than trying to use speech recognition to open applications, etc.) The camera is probably very similar to the $20 one at Best Buy.
Re:$7k?! (Score:2)
7k basically buys you a good camera and a full computer.
eye tracking and head tracking and etc trackign solutions are old as fff..... and it's not expensive because of the "enermous r&d", with that logic anything is expensive - and the costs for the r&d are largely just something they agreed on by themselfs. the market is small because there's not that many uses where it's actually good now, the use for the pointing device if you can use a normal pointing device is very limited.
Re:$7k?! (Score:2)
the market is small because there's not that many uses where it's actually good now, the use for the pointing device if you can use a normal pointing device is very limited.
Yeah - small market (handicapped computer users) means high prices. Another reason why it costs $7k.
eye tracking and head tracking and etc trackign solutions are old as fff..... and it's not expensive because of the "enermous r&d"
If they actually have it working smoothly and reliably, then no, it isn't old. Try some motion tracking in After Effects or even Mocha's tracker and after seeing the ridiculous limitations of it you'll see how far behind the field is. Motion tracking is _not_ a developed field. It's still very much in the wild and it takes a certain type of engineer (i.e. high salary) to wrestle with these types of problems. And yes - the development costs _are_ enormous relative to your typical "hey let's write a webcam driver using the textbook technique that everyone on the planet already knows" projects. Sure there is plenty of research in the area but only the most recent stuff is actually applied research. (how would you have applied the research prior to the invention of computers and digital cameras?)
That seems to be cheap actually. (Score:2)
Being that these guys apparently have a number of products in this market already, I suspect they already know what the price levels are. (I bet a lot of their stuff is paid by insurance in whole or part too)
I'm not saying it's not expensive as heck, but that's how things are in those low-volume/high-margin markets. You know, an PCR machine for a biotech lab isn't a heck of a lot more advanced than a digital toaster, but the price difference is an order of magnitude.
Re:That seems to be cheap actually. (Score:0)
These sort of things are always expensive because the market is small. Hearing aids are still ridiculously expensive despite being relatively simple. If normal people had use for hearing aids the market would grow and prices would fall. The good news for the stroke patients are that normal people seem to want to use advanced sensor bars and maybe even sensor arrays for things like games.
OpenGazer (Score:4, Informative)
Re:OpenGazer (Score:5, Interesting)
What could possibly go wrong? (Score:0)
DON'T LOOK AT THE AD DON'T LOOK AT THE-- Arrgh
This sentence is to get around the caps filter.
Free vs & grand (Score:-1, Troll)
Re:Free vs & grand (Score:0)
How many more times as AC are you going to post to spam your link?
Re:Free vs & grand (Score:-1)
FPS games (Score:2)
this should even the field between the noobs and pros with "skill". Can't wait until someone produces a server mod for [insert favorite fps here] that dis-allows this device.
What about the pointer? (Score:3)
Darn that new eye-controlled mouse! Everywhere I try to look on the page, there's a pointer in the way!
Re:What about the pointer? (Score:3)
I'm disappointed that we haven't seen anything beyond a simple "directly control mouse pointer" implementation. There are so many more interesting things one could imagine. For instance a "focus follows eyes" like "focus follows mouse" would be interesting. (Not for a general user, mind you, but as an option for people who have a certain kind of workflow.) Something else I would love to see is a "smart snap-to". For instance if I'm looking at a given widget, and I move the mouse pointer from far away towards the widget I'm staring at, it could immediately "snap" onto that widget, so that I can easily click/interact with it. It would only work if calibrated correctly so that it isn't triggered during normal operations (e.g. I'm moving an object in a program, my eye focus is fairly close to the pointer, so it shouldn't snap annoyingly) and only exists as an 'accelerator' during big pointer movements. It could actually be implemented as a sort of region around my focus that has 'slower mouse acceleration' so that the pointer sorta gets trapped in that area (effectively creating a Fitt's Law "big target" where I'm looking), probably coupled with some context-awareness (e.g. only active when I'm looking at UI chrome, not movable elements within a program or the canvas of a drawing app...)
The devil is in the details, but I think there's a way to implement it where it would be helpful in many cases (the pointer magically moving where you wanted it to) without screwing up other cases (when you want to delicately move the mouse). I'm sure there are lots of other neat ways that eye-tracking and mouse movements could be ingeniously combined (when my eyes are searching around frantically, make the mouse pointer glow so that I can find it?)... a simple control of the mouse pointer doesn't sound great.
(One problem with the things I'm proposing is that they are so subtle and such small gains in productivity that they certainly wouldn't justify a cost much beyond what a webcam costs...)
Re:What about the pointer? (Score:0)
I've worked with this company and their technology before, and believe me when I say there are more applications than simply controlling the mouse pointer. However, just like a touch screen eye control is something that works best if the interface is designed for it. As such mostly specialized applications benefit from it, and most of them are aimed at people who have a hard time using a normal computer interface.
That being said, there are programs for the rest of us as well (although this is mostly a work in progress) that benefit greatly from eye control. You can probably imagine a few times when hands-free operation of a computer would be beneficial.
Re:What about the pointer? (Score:0)
I'm disappointed that we haven't seen anything beyond a simple "directly control mouse pointer" implementation. There are so many more interesting things one could imagine. For instance a "focus follows eyes" like "focus follows mouse" would be interesting
No... just, no.
I use Focus Follows Mouse in KDE and that's the way I like it but it requires a specific mindset to use correctly; essentially, if I have several windows and I'm coding or writing a document that uses information I'm reading from another window then I do NOT want the keyboard focus to shift away from the editing window just because I glanced at the other window whilst typing.(*)
(*) Sure, sticky focus that doesn't change until you stop typing but how long do you wait? What if I pause to think whilst staring at a random location on the screen before I resume where I was, having the stick time less than 10 seconds will just break my concentration but at that length of time it would be too slow to be useful in the general case.
Re:What about the pointer? (Score:0)
The Tobii-Lenovo laptop has Tobii software with mouse pointer behavior just like you describe. Having run it for a couple of months on my work computer, I get annoyed over the slow pointer movement when I use my home computer.
Just like you say, naive implementations with the mouse pointer just following the gaze point (or aiming using your gaze in FPS games) are close to unusable. Tobii is very aware of this and are working on concepts for using the gaze information in smart, non-intrusive ways, to enhance interaction rather than completely change it.
Re:What about the pointer? (Score:0)
Why would you need a pointer? If the eye tracker is accurate enough, there should be no need for a pointer.
Re:What about the pointer? (Score:2)
Have you never heard of ShieldCursor()?
The cursor goes away when you start typing. Also, I don't know about this one for sure, but similar ones in the past required you do hold modifiers to indicate you wanted it to follow your eyes. (Yes, I think having to hold modifiers+ use eyes is still optimally way better than a trackball, which I use instead of a mouse.)
I would love one of these. (Score:1)
The thought of being able to work/play grindy mmos without having to agitate my wrist sounds wonderful. (But bad too, because I might get addicted again)
$7k is absolutely unreasonable though, hopefully in a few years and with a competitor or two it'll drop down to something your average guy can buy.
wont work (Score:2)
Trolling. (Score:2)
Do they assume Apple users only have one eye?
Cheaper spam (Score:0)
with SmartNav [naturalpoint.com].
No retinal risks either.
Looking ahead? (Score:2)
Not convinced about the practical use for this, your eyes are generally working one step ahead of your hands.
Let's say I'm playing Bejewled in timed mode, where you need to make moves as fast as possible- my eyes are already looking for the next move as my hand makes the move I've just found.
Same goes for browsing the internet or many other tasks, where my eyes are looking at something else while my hands move the cursor to a link or, say, Next Page button...
Ask any guitar/bass player, they're not looking at the note they're playing, they're looking at the next note their hand is going to move to.
Having the cursor track the eyes would significantly slow down a power user I think?
I've seen another used for ALS and the like (Score:0)
I've seen some demos on the local news from a Vancouver company with similar technology. The news program was on neuro patients like ALS and MS. The eye tracking seemed to work well and if I recall it was cheaper than 7k.
Mirametrix.com. They have some demos up on youtube.
SmartNav optical tracking (Score:1)
Natural point tracks head movement and has for almost a decade.
Seems much cheaper and simpler than $7000.
http://naturalpoint.com/ [naturalpoint.com]
Re:SmartNav optical tracking (Score:2)
Overkill and not a mouse (Score:-1)
A webcam with the right software can do eye tracking just fine and will hardly cost $7000. And if this is a mouse, I suggest updating our terminology:
trackpad = flat mouse
trackball = upside-down mouse
stylus = stick mouse
joystick = penis-shaped mouse
touchscreen = invisible mouse
Not its name (Score:0)
Its name is Kunta Kintei.
can it be used ... (Score:2)
... by Stephen Hawking [wikipedia.org]? I'd be impressed if so.
Re:can it be used ... (Score:0)
I'm glad you provided a link, as a Slashdot user I would otherwise have had no idea who Stephen Hawking was :p
Great (Score:1)
Now I can get Carpel Tunnel Vision.
iAimbot? (Score:0)
Uh oh, it's the ultimate FPS hack!
1. *eye twitch*
2. Click.
3. *BOOM HEADSHOT!*
hax (Score:0)
Great, now people can shoot you in FPS just by looking at you.
This has been available for a LONG time! (Score:0)
These guys marketed a good eyetracking system in the 1990s: http://www.eyetechds.com/
Initiate actions...how? (Score:2)
FTFA
...after installing some software, users will be able to control various onscreen actions by looking at, blinking at, or fixing your gaze on an object.
So what, two blinks is a double-click? A wink is a right click? What about when our eyes are just plain dry?
This could get disorienting real fast...I'm not saying it's not good for some special circumstances, but IMHO people just aren't as consistent with their eyes as they are with their hands. Even while typing this, I'm looking around my desktop, checking my notification icons, etc. (ooh, I have mail) It would be disconcerting if my cursor were jumping right along with me, and downright frustrating if it shifted focus off of the text editor because I blinked at the wrong time...
FPS gaming (Score:0)
I'd love to play counterstrike with this.
Yay (Score:0)
I can finally set my window manager to "focus-follows-eyeballs"
Free face-tracking software for Windows (Score:1)
We do a similar piece of software that tracks your face movement: it's free, and you can get it from http://www.facemouse.co.uk/ [facemouse.co.uk]
Install (Windows only), run, position your head facing ahead at the webcam, and then move the mouse around by turning and raising/lowering your head. There are two versions, one that click automatically when you stop moving your head and one that doesn't (so you can use another dwell program of your choice.)
Jason Becker (Score:1)