I’m working on installing an LCD panel behind my acrylic, as described in the last paragraph of Jeff Han’s siggraph paper.
If you create a frustrum where the large side is the acrylic, the back smaller side is where your camera is, you can include flourescent lights, (since they dont have much IR) to essentially “relocate” the backlight.
I’ve already dismantled a cheap LCD screen. I put it up to the camera-side of my acrylic and shot some IR tests. You can see the images at my site.. ( http://www.mikepkes.com ).
just want to add that this is also a solution for not having to use a projector, because the image is on the lcd instead of on a projection screen where it’s being displayed by a projector
this way you don’t need alot of space behind your screen for the projector or having to place mirrors
downside is that you’re stuck to the size of the LCD you can get your hands on
Hey, looks like you have a nice looking setup going there, I’m interested in doing the same thing. I was wondering if you could elaborate on this frustram though, I can’t quite figure that part out. My idea was to use an LCD
with the back light removed, but shine a white projection at it(would need to be much less powerful than a video projector i assume.) Have you tested with the LCD powered on yet? I would think that when the screen is on
and white the blobs would be way brighter than when you are looking at black. I don’t know how LCD respond to IR though.
I am interested in doing the same. As others have said here, some tips and experiences would be nice.
I just have a question:
- I have read lcd screens have lights at the side that iluminates and those produce a lot of IR light and can cause the system to not being able to detect the fingers (or whatever you use to touch the screen). Also have read that the lcd that uses fluorescent dont do that. Would be nice to know what lcd you used and if you (or anyone) can give some tips on what lcd to get.
Also if anyone can tell me a webcam that has the maximum angle, so the distance can be shorter? Thanks
I am a complete newbie to screens so thanks for the help.
Ok, as I am a complete newbie to the design of an lcd screen I have been looking around the web for info. I have seen a page where they dismount a lcd screen and it was pretty instructive. The article is this one: http://www.bit-tech.net/hardware/2006/03/20/how_crt_and_lcd_monitors_work/1 Check sections 5 and 6.
This lcd screen is lighted from the sides, and doesnt have backlight. Basically any lcd screen is composed of the display matrix with polarizing filters on each side. As people said in this post, those panels leds infrared light through. But there is also some more panels behind that. They are some kind of material that difuses evenly the light that is being send from all the sides to go into the display. Its a withe panel in the web page that I pointed before. This is a good thing because you get rid of the backlight and have more room for cameras or other electronics you want to use. And also less heat. But my question is: does this difussing panel let infrared through?
Cause if it doesnt, backlight is needed. In this case, where the difussing panel doesnt let IR through and backligh is needed, how do you send the light even to the whole display panel. Cause if you just use some bulb the center will get more light than the sides. Does this affect what you see or does the display panel compensates this?
Hopefully the difussing panel will let IR through but in case it doesnt can anybody answer my questions or at least point me in some way so this things can be solved.
Ok, so today an argentinian guy came to the irc asking for some questions, and he is planing to do an lcd ftir as well, and as he had allredy dismounted some lcd screens I could get some good info from him. Thanks Esteban
So, he said IR light gets very difused when it goes through the difusor (indeed) and you get a bigger dot than it actually should be. Also the light (including IR) of the backlight gets reflexed on the difusor and goes into the camera, and that means problems. He told me you get similar problems with side ilumination.
So it looks like to do this you have to get rid of the difusor panel. But he told me that if you get rid of it, you can really see where the lamps are when you look at the images. A posible solution that someone pointed in this forums is to use leds as backlights. I am guessing that if you use enough leds you will get an even backlight ilumination. Anyway this is just theorical thinking, and a test should be done to see is this is doable.
So, first Ill have to build the detection mecanism and after that I will dismount an lcd and will play with it. Any ideas or suggestions are welcome.
just to echo hugolp, for the people who are reading up on this subject, the idea is to remove the backlight and the diffusor and illuminate with some sort of raster of leds for a good spreading of the light
or how about placing the diffusor and the light in front of it
Yesterday my friend and I started working in a rear/diffuse illumination multitouch display. We successfully dismantled an Acer 1916 panel and have the bare LCD. We’ve set that aside for now, and are starting to test the actual finger tracking.
We haven’t made our own LED array yet, but are instead using the infrared illumination from a Q-See B&W Infrared Camera. We’re using the Microsoft Lifecam vx-6000 with the IR filter removed and two layers of developed, exposed color film in its place.
The initial results have been disappointing. The Q-See provides sufficient IR illumination, but we were unable to make the fingertips provide sufficient contrast to be picked out as being distinct from the rest of the finger (or even hand). We were just using bare acrylic, as my understanding was that a “compliant surface” is only necessary if one is using FTIR and not rear-illumination.
What am I missing here? mikepkes, what are you doing to get your fingertips to show up so distinctly? Is there a particular setting of the touchlib filters, or is there some other component in the mix? One thing I was considering was that perhaps the IR illumination is too bright, and I need to put additional filter material in the Lifecam.
That LCD demolition link that Hugolp provided looks an awful lot like my LCD, the LG LM170E01-A5. It was fairly straightforward to get to the TFT display, but it didn’t want to come apart easily. I dont have an FTIR built yet, so I put it back together. The manual says the insides are very delicate and it is not designed to be dismantled. I don’t know how accurate that is.
with diffused-rear illumination you need a translucent material placed on top of your acrylic/screen. Without one, IR light will shine through the screen and reflect off your finger, hand, arm, body, etc…
the “compliant surface” in FTIR refers to a specific quality that a diffusing screen must have for FTIR, this specification does not apply for diffused illumination (DI)
with DI, you have potentially more options for a diffusing screen: mylar, tracing paper, sandblasting, sanding, spray-on frost, etc… BUT, since you are using an LCD panel, you might want to use tracing paper since it doesn’t blur. the only image degradation should be that the image is around 50% less bright.
as for the LED array, I dont know the specs of your LED’s or how they are used, but you dont want the LED’s to shine directly onto the surface and bounce back. Place your LED array to the side and to the ground (get maximum distance away from the screen). in theory,you want to illuminate the area indirectly because the hotspots from LED’s are really strong (you can not diffuse the LED source). ideally, your camera shouldnt see hotspots, in practice, as long as the LED hotspots are not shining in your interactive area, it’s ok because you can crop that out.
Excellent, thanks for the tip. We got some drafting paper and that seems to do the trick. Next questions:
1. Should we use the diffuser sheets, diffusing cell, and side illumination flourescent tubes from the LCD itself, or get rid of that stuff and just install our own flourescent lamp alongside the IR light source?
2. The diffuser sheets, cell, and bare LCD crystal panel are all glossy or shiny. How do you prevent the camera from imaging a specular hotspot of the IR source? (Or is our source just not diffuse enough if we’re having this problem?) Should we perhaps put the drafting paper between the LCD and the IR light, instead of in front of the LCD?
3. Is acrylic even necessary? What purpose does it serve except to protect the surface of the LCD as I move my greasy fingers all over it? If this is the case, why not use something thinner like a big sheet of transparency film?
