Slashdot Mirror
Gaze Gaming Tech Promises Faster Eye-Controlled Interaction
NewScientist is reporting that further research is progressing on new types of user input devices. Specifically, "gaze gaming," a technology that promises faster interaction using only your eyes. Currently technology for sight-based interaction is far too slow for practical applications in things like gaming. "Eye-gaze systems bounce infrared light from LEDs at the bottom of a computer monitor and track a person's eye movements using stereo infrared cameras. This setup can calculate where on a screen the user is looking with an accuracy of about 5 mm."
I'll bet this will ruin my eyesight for half the price of Lasik.
Two kids trying to figure out Wild Gunman in 2015:
"You mean you have to use your hands?"
"That's like a baby's toy!"
Looks like we're not far off, considering we have 7 years to go.
While your finger sits on a touch sensor (unmoving, relaxing) your eyes act as the mouse curser. You blink to click. Perfect interaction.
I always wondered if you could do more precise gaze detection by looking at a person's retina. Could you detect where they were looking on the screen precisely enough to eliminate the need for a mouse cursor (say, within one character space)? How large is the area of sharpest vision? /frank
And the worms ate into his brain.
at least for FPS.
Otherwise, mostly a Gee Whiz! tech, though I suppose it could have useful applications for the disabled. But I wonder if we won't see wrist-based Repetitive Motion problems transferred to increased eyestrain...
Freedom isn't free; its price is the well-being of others.
And then stick with your trackball. Thx.
5mm is half a centimeter. On a 24" widescreen (61cm diagonal) you've only got (2*61) 122 steps, there are 2203 pixels (diagonaly, I know it's wrong). This means you've got 122 steps of 5mm to cover 2203 pixels. Not accuarate enough, by far.
Improve to 1 or 2 mm then we will talk.
Greetings, a no-life gamer straight from his basement.
It would be really useful to be able to move the cursor only by looking at the point on the sceen I want it to be. That could save my wrist from carpal tunnel syndrome and it could also incement my productivity by making the pointer go quicker to where I want it to be. I hope it will have pixel accuracy, but even if it does not, I am sure, time a few years, it could become the perfect input device.
I just don't trust anything that bleeds for five days and doesn't die.
blink enough when I play games. This could only make things worse
One of our competitors trademarked the term "hypothesis". From now on, we will call them "boneheaded ideas".
This means Nintendo's next generation of console will be called the sii?
There is going to have to be a very accurate system of "disabling tracking". I mean, take the FPS example. How often am I going to be spinning around when I glance down for a quick ammo count?
Or, if you look up to check your HP/MP in an MMO, will you be randomly changing targets, or worse, disengaging them to move?
Its an interesting idea, definitely useful for somethings, but it shouldn't ever take the place of a mechanical pointing device like a mouse or trackball.
If sharing a song makes you a pirate, what do I have to share to be a ninja?
I am always scanning the whole screen, and just going back to where my "focus" is.
Think about it, even if you are playing a fps and aiming at something, you are still looking all around for other threats.
Or if you are coding, writing, etc. you are often looking at other reference materials outside of your current focus.
My lips don't move when I read, and I can type without looking at the keyboard OR the screen...
I don't think of eye tracking as an ideal form of input, especially not for FPS games.
What would you even use it for? Moving? Aiming? It'd be a poor choice either way, and you'd still need other input for all the other stuff you need to do.
It would be surprising if it ends up having any practical applications at all.
Some of us need glasses just to see up to the screen. How will this work with an additional semi-reflective layer interspersed?
Sig Battery depleted. Reverting to safe mode.
The pioneering of this fledgling technology will be done by the porn industry. The implications!
For single-player games, this device could possibly enable some sort of selective rendering technique, where the objects sitting at the focal point are rendered in much more detail than the periphery.
In Soviet Russia, TV watches you!
I've calculated my velocity with such exquisite precision that I have no idea where I am.
I'm cross-eyed you insensitive clod!
TrackIR seems a lot more useful- it captures your head motion and creates an effect like looking out a window. I am surprised that it has not found applications beyond the hardcore sim crowd.
I'd be curious to see what might happen in some sort of "interpretative" game if attached to someone experiencing REM sleep.
You know, for science.
While I can see (no pun intended) good possibilities for games, it may cause a few problems if it were implemented in a desktop environment (as a few here are suggesting). I know at the moment, with the firefox tab-mix plus "hover the mouse over the tab to select it" feature that while it is useful, it can also be annoying if I don't park the mouse properly, as I can suddenly switch tab when doing something else. I could envisage the same problems with this. Unless it was easily deactivated a stray glance or blink could be shuffling me around my desktop environ.
For games though it presumably would add another immersion layer, which can only be a beneficial thing.
Now we can farther yet remove gaming from pesky physical activity. It's no longer even a matter of good hand/eye coordination: just good eye coordination.
This seems like a good idea in theory, but in reality we rarely keep our eyes fixed on any point with all that much precision. Our eyes are always shifting around to get a bigger picture of things most of the time, even when we're trying to hold a steady gaze on something. Trying to precisely control a game, or anything else, with one's eyes seems to me like much more trouble than it's worth.
Slashdot Burying Stories About Slashdot Media Owned
...will happily run out to Gamestop and purchase the latest version of "Bugs, Fishes, and Shiny Things" for our feline overlords of 2015.
Other than that, and possibly use for the handicapped, I don't see a use for this tech in games/applications. It'll be as reliable and annoying as voice recognition software.
Simple head tracking would be more useful. Tracking eyeballs fixed on a 2D plane will surely have horrible calibration issues, problems with multiple people looking at the same screen, problems with cats, glasses, contacts, glass eyes, monocles, etc.
Even if it works out great, the bottom line is that I don't WANT this technology. Certainly not for my general pc/web usage. Maybe for a game or two, but it's bound to be a novelty more than anything.
Then again, the prospects for porn are...titillating.
Forget "the great equalizer". With my lazy eye I'll be staring up at the heavens spinning in circles all game long ... that is all game until my head gets blown off repeatedly.
Look at area,
say "click"
or
say "nudge [right|left|up|down|north|northwest|west|..." , then say "click|pow|go|do it".
She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
I remember the pepsi commercials back in the late 70's/early 80's.
They tracked where guys were looking and it was not at the product.
In fact, they frequently didn't remember the product.
Very popular commercial of a girl exiting the water in a little suit holding a can of pepsi.
She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
Why replace the (mouse | right stick) in an FPS? Add this as another input.
Keep the left stick or WASD keys controlling footwork, and the right stick or mouse controlling the point of view, but add this to allow interacting with some place other than the center of the screen.
The only problem I see is that, the way it works right now, you can spot a target and move the mouse so the target is at the center of the screen. Once you've got it there you can shoot at it while looking around the periphery of the screen for your next target and sort of focus half on you current target and half on finding/keeping track of other targets. If your aim follows your eyes, you lose something... no longer being able to shoot in one direction while taking a little of your focus in another.
*Excessive cursor movement detected. Add selected*
Sore eyes from too much Gaze Gaming cursor control? Buy Visine!
"My eyes! The goggles do nothing!"
"Dude, you can't control Mario if the Sii can't see your eyes."
"Oh..."
Try chess or Go.
Yes, and eventually they'll create a console that can read your intentions without any conscious effort on your part at all - all you have to do is exist.
This console will be called the Bii.
NewScientist, hmm shouldnt it be OldScientist? Take a look at www.tobii.com .. been around for a while ,)
Dude, this is exactly how they almost took over the Enterprise that one time when Wesely came home from the Academy on vacation. I wouldn't trust it.
Cogito Ergo Sum
Great. Now my beautiful better half doesn't need to try to get me to acknowledge her repeated pleadings for my attention while I game on. Now all she needs to do is put on that top I really like and stand at the edge of my peripheral vision.
Maybe if I rig a set of blinders...
I use irony whenever I can, but my shirts are still wrinkled...
Quickly summarizing a few comments I found interesting from the threads (instead of quoting them all);
- potential for gaming/military, maybe training sims, bad for desktops
- third input, as in mouse/gamepad assistance
- handicapped assistance
- head tracking benefits
- potential issues with glasses/contacts
- 'aiming' accuracy based on size/distance of screen
And then an additional comment:
So what if the thing was added to a helmet, with HUD. The fixed (configurable) distance between eye and sensor should increase accuracy. That would also allow for head tracking (wii style). My Razer Diamondback/Tarantula) is very generously configurable, and only uses 3 USB ports. Something like this may take 2, which could be prohibitive. It might be best with its own PCI/PCI-e interface. I'd imagine if it was a 2 way device, it may have even more practical uses beyond gaming.
----
I'd have to remove my tin-foil hat to make room, of course, and then the black helicopters would find me.
Be careful about reading this headline aloud. People nearby may get entirely the wrong idea of what sort of sites you visit.
When someone says, "Any fool can see
for virtualty (as opposed to virtual reality which requires you to build in everything.)
By making these things into eyeglass frames and using geo-positioning you can interpose a reactive layer between you and whatever you're looking at.
This means great possibilities for 'non-intrusive' gesture (or bluetooth device triggered click,) capture and subsequent playback for whatever you're interested in that you're looking at.
What I could do with something like that...
MSBPodcast.com The opinions expressed here are my own. If you don't like 'em... Think up your own stuff.
Imagine using this feature to render highly detailed 3D images only where the user is actually looking. The peripheral vision is almost useless, the actual area that does most of the seeing is in the center of the retina. The brain fills in the blanks and keeps a mental image of what you are seeing that gets updated through rpid eye movement. With a smart setup this could be translated in a huge screen that appears to have a large resolution in every direction while keeping the processing power requirements still accessible. The mouse is a great interface, I don't see any reason to replace that just yet.
Will it work for Marty Feldman? http://www.imdb.com/media/rm3708590080/nm0001204
We need a "+1 -- nice sig" moderation.
John Carmack spoke about this at QuakeCon 2007, not from a controlling in games perspective but rather from an optimization point of view. The eye really only focuses on one point at once; everything else in the field of vision is blurry. His idea was to only make what the eye was looking at on-screen high-detail, thus allowing much more detail in scenes.
If there's anyone I hate more than stupid people, it's intellectuals.
It's probably not that great on its own for able bodied people, but as one of many inputs it could be quite interesting. It could be used to subtly alter camera views towards what you find interesting, I suppose. This sort of stuff is already in use to generate "heat maps" of where people are looking when using a computer. Like where on a webpage you look, or in game, etc. It'd be useful to know what sort of things people don't care about, or how distracting something fading in / out might be to the task at hand.
It's also usable with dasher as a fairly fast text input for the disabled.
I Browse at +4 Flamebait
Open Source Sysadmin
At least they won't have supply problems with the Si. As an artifact, it keeps duplicating itself until you end up carrying too many of them around to manage...
But how will they combat 'piracy' then? And does anyone here understand ADOM jokes?
I work in a cognitive psychology/psycholinguistics lab where we use an eye tracker to track gaze duration while subjects read.
First, you have to keep your head very still while gazing through the eye tracker. It's pretty much a device with a forehead rest, an adjustable chin rest, and an angled piece of yellow glass in front of the eyes. The camera projects infra red light from above and reflects them off the mirror to the eyes to track their movement. If the subject takes his head out of the tracker, even for a second, we have to discard that trial and recalibrate the machine. Because the subjects can't move the head out of the device because of calibration issues we have to give them short breaks every ~5-7 minutes. Of course this break is followed by a 2-3 minute recalibration and validation sequence.
Its also notoriously finicky about even tracking people's eyes. Girls wearing heavy mascara and some people with glasses hardly work at all in the machine. Sometimes people with contacts or uncorrected vision can't even be tracked.
Until substantial progress is made with devices like this no one will be using them outside of research work.
followed by that last model to which nobody paid really attention, because everyone was fed up with the marketing.
I think it was called "Gii".
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
I don't have exact number handy but perhaps a crude 320x240 (antialiased) resolution might be enough for peripheral vision and the 1600x1200-equivalent-DPI-wise could be kept only for the region the eyes are currently looking at.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
Ohh great, installation programs that require me to read every charter of the end user agreement before installing my games.
I would love this technology, every time I look at the screen there is a cursor exactly where I'm looking.
Imagine the advertising policy's that could be implemented. Hotmail will now require you to look at all the advertisements before checking you email.
It will not be enough for hardcore keyboard+mouse players. I'm not that skilled on FPS yet I have found myself selecting targets with my eyes, then moving my hand to click on them while my eyes start looking for more targets at the same time.
just my two clicks...
Most games have information on the screen that you'd like to look at, but not necessarily shift your viewpoint to. Imagine looking at your health bar in an FPS and having your POV shift to the bottom left every time. It's probably useful in other situations where the perspective isn't locked to a cursor.
I think this would be great as an alternative to a mouse for many information-based applications (ie: web browsers, word processors). I do a lot of programming and it's often annoying to have to move a hand off the keyboard, to the mouse, select what I need to, then move back to they keyboard. Most apps simply don't do hotkeys well.
As for games, probably good as an additional input. But it would probably be too straining and often used improperly. Like in FPS, when you're shooting someone you're often glancing around the screen to see other opponents. It would be bad to have the crosshairs locked to the focal point. Probably would work well for strategy games.
... but soon, they can frag.
Should still give better control than a laptop touchpad, at least I'll know why my uncontrollable clicks are happening...well, until they load up the eye tracking with command gestures and special click areas that are enabled by default, anyways.
"When information is power, privacy is freedom" - Jah-Wren Ryel
...so in my case, I've tried.
It's ridiculously easy to play as long as the ball is moving at a similar speed to the paddle. Once it goes faster it becomes nearly impossible, because every time you look up at the ball to track where it is, you undo your work trying to anticipate the ball's position.
Also, as a photo geek, I can state pretty emphatically there are plenty of times where you're not looking at the primary subject when you hit the shutter; many times you're waiting for the perfect background moment to match your foreground, or for the idiot gawking at your photography to finally be obscured by something in the background moving in front of them.
Is it workable? Probably, but the interface isn't as cut-and-dried as a lot of these articles like to claim.
In any FPS gaming situation, you need to be accurate within one pixel, and you need to be lightning fast on top of that. 2mm can make a difference in score equivalent to the difference between a master gamer and a drunk frat boy playing the game using his asscheeks. Case in point: Metroid Prime: Hunters on the DS. When you meet a 4 or 5-star player, your entire understanding of what is humanly possible may change. When you get to the level that can put up a good fight against them you'll definitely understand what I'm talking about.
"When information is power, privacy is freedom" - Jah-Wren Ryel
I was thinking at the OPENGL/drivers level.
- There exists OpenGL implementations which could get OpenGL command and redistribue it accross several servers on the network, each rendering a part of the diplay.
- Poorman's Anti-aliasing on the early GeForce / TNT was done by intercepting OpenGL/DX commands and rendering it at a different resolution in a buffer and then subsequently smoothly downsizing the buffer to the screen.
- Current nVidia drivers produce stereo for games that don't support it, by intercepting the commands and rendering the scene twice, each time offsetted to acknowledge the position of the two virtual eyes of the player.
I was thinking of something along the same lines.
Game works as usual, games sends commands to graphic card to render the scene, the modified opengl library intercepts the commands and distributes them to several jobs each rendering a different part of the scene in a buffer using a different resolution. Then those buffers are composited to the screen (the currently looked-at buffer is copied as-is. the lower-res buffer are bilinearily zoomed during copy).
Of course, this kind of approach works much better with tile-based renderer than with classic graphic cards, because they segment their output in a similare fashion already.
One advantage is that jobs could be spread accross several GPUs. (Think CrossFire/SLI, but not in Alternate frame rendering mode, but Split screen rendering mode, where some tiles of the split screen aren't rendered using the same resolution as the others)
*BUT*
The main problem with that kind of approach, which is also a problem shared with SLI when in Split-screen, is that modern games (such as Crysis) tend to post process a lot of their output, meaning that the software spends a lot of time rendering to buffers different parts of the scene, and only renders to the screen when compositing and applying the filters. (The different kinds of depth- and wholescreen- motion blur, HDR and gloom, etc...) That's one of the reason why games like Crysis are so power hungry and scale badly in SLI beyond two-GPU configuration.
For those games, I could be easy to do the variable detail in the engine it self (split the scene to several varied-resolution buffers when rendering the main part of the rendering before the compositing), than in the drivers (No point in splitting the scene when drawing the screen because, usually, it's only 2 big triangles used to copy the main buffer to the screen, with a tons of post processing shaders running on them. What should be split is the various buffers before, but it's hard for the drivers to guess which buffers are going to be relevant)
Or, the game engine will need a new standart to flags which buffer are only some small details, and which buffer are actually the meat of the scene. This sound complicated, but on the other hand will help *A LOT* the organisation and work-unit distribution for SLI.
*ALSO*
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
Imagine the popup ads that would undoubtedly take advantage of this technology...
You know those little floaty see-through things you see in your eyes, and they always jump away when you try to look directly at them? This would be just like that, only totally opposite. And shaped like wing-wangs.
One man's constant is another man's variable.