Awesome interface

… but there are a very poor info about how it it done :’(


A student of mine is working on a similar topic.

the fact that the video is shot in darkness is also suspiciuos :))

it might not work well with ambient light

Yeah, I can’t see how its working, too!
Especially because the emphasize that it’s low-cost.
Well there is a sketch but I have no clue how they detect the light. Is it just a row of light sensors on each border of the frame?


I’ve read about it weeks ago and with my collective I’m gathering all the information I can. I think that is simple as it seems, a plexiglass square covering an LCD monitor surrounded by infrared LEDs, using the FTIR technique we can track multiple points on the sreen. Using libraries like JMyron for processing we can then capture those tracking points and make Processing draw something… Bartolomeus is the RESOLUME developer (resolume is a VJing software) and on its forum told that he could make sometinhg working with this kind of structure… very interesting topic

yeah, but how exactly do they implement FTIR?
I’m pretty shure now, that it’s just a beamer behind the screen and beneath that is camera to detect the finger movement. As massimo allready noted, there’s no room light. I guess the contrast of the FTIR image would be too low in this case.

I think that its not as simple as shown in the picture, but maybe tracking finger shadows in an infrared backgorund may be factible.

I did take some photos with my web cam of the shadows produced by my own fingers over a paper sheet attached to a glass and I achieved very recognigzable patterns of dark circles. However it may be difficult to create an algorithm able to detect them. I´m quite interested in this issue, so I will take a look to those libraries for processing you say.

It would be nice having an “open” version of this technology ¿isn’t it?

I think that infrared light is used because it is not “visible light” and so it does not make noise with it… so my idea is that visible light (from a light bulb) will not interfere with infrared light (and the dark room is just to make more visble the LCD screen under the plexiglass) but this is just my idea (and now is 2.00 am, so…). As an answer to quique, I would like to say that maybe is not necessary to write a pattern-detecting algorithm, as Jmyron already implements them. If it will be not sufficent, maybe we can use Eyesweb (, but it’s use is out of my knowledge… :-/




at my work some people developed this, based on the technology of the electronical instrument “theremin”:

the big advantage instead of using optical tracking is that you will have no problems with changing light etc.

another interesting inteface for multitouch is this:

but it is very high prised :slight_smile:
when working with Processing have a look to

cheers fuxle

Apparently Philips already made a commercial version of these:

…kinda weird since i read somewhere that apple is holding the patent for that technology. well… competition makes it cheap hopefully.

ambient light does not seem to be a problem, they might have used some simple infrared filter on the screens surface to keep unwanted infrared interferences outside:

"While the concept of a multi-user digital table is not new, previous solutions have utilized complex arrangements of overhead cameras and dimmed lighting that detract from the user experience. The Philips Entertaible, however, is based on a series of infrared LEDs and photodiodes discretely mounted around the perimeter of an LCD screen. It requires no special lighting conditions or other equipment and is entirely ‘hand’ operated by touch alone. Entertaible can simultaneously detect dozens of objects, including fingers. "

…it would be nice to know whether it is possible to build that thing at home with no superbig effort :wink:

There’s this whitepaper on touch screen technology from NEC… not that recent, it’s from 2001. Nice overview, though.

You might have seen it on Tom Igoe’s site before, but his link is broken.

If you are interested in working out how multi-touch screen using FTIR, I’ve posted some more people doing projects in this area