Hi everybody,
first post to submit an idea.
A couple of years ago I was working on a project that used some equipment to measure the topography of a surface using fringe deformation ( not sure about the word “fringe” in english though). The principle is as follow: a projector displays fringes (basically parallel lines horizontally or vertically) on an object. The fringes are actually moving in phase over time so that you raster the object constantly. A camera look at the object and an algorithm look at the deformations of the lines on the surface. This gives basically a 3D image of your object.The method can be actually really precise. As an example it was possible to look at the circuit on a silicon wafer so precision can be very good. One applicatoin of this is also to measure the deformation of an object over time very precisely, that is actually what can be used in MT.

Now two possible applications of this for MT.
1- “simple” case: by compositing the fringe with your normal desktop image on a projector and displaying this on a surface, a very little pressure can potentially detect the deformations and give you blobs by focusing the camera on the “screen” or directly looking at the deformed fringes on your fingers (probably more ambient light sensitive).
2- more complicated: using a LCD screen and compositing the image with the fringes and focusing the camera THROUGH the LCD to the surface you touch.

Technically it reduces the problem as we do not need any IR lights. In the first case the projector problem remains. In the second case however because the screen itself produces the fringes ( assuming it rasters it quickly enough so that it is not visible to the eye) , the tricky part is to be able to focus the camera though the lcd onto the surface that deform.

just my 2 cents as I walk around

Hi,
I hope you read this post, even if it is several month late.
I think there is some potential here.
But lets give it a twist: lets project a grid of IR light directly on a deformable surface (as lycra, elastic fabric, etc.)
So, you don’t need to rely on high frecuencies to overlay this grid, and you can use a regular IR sensitive camera.
This also eliminates the need for glass or acrylic, making it very cheap. The fabric also is used as projection difusser.
The IR grid can be made using LLP technique or similar (with cylidrical lens)
Well, is an interesting start for an interesting idea.
Best regards, and lets keep spining this.

This was actually done at siggraph this year. A grid of lines on a object deformed when pressed and a camera picked it up. It’s not a bad idea, but pressing on something hard is much easier and natural compared to when I tried this out. Also, you can’t really use an LCD for this technique. I think there would have a be a real big benefit (other than cost since acrylic/glass doesn’t really cost THAT much) to really make this a strong and robust technique. I’m also assuming it would take heavier calculations to process how much the mess deforms.

Wow, that was fast!, already in siggraph!
Well, it certanly will not work with any solid surface (as an LCD) behind it.
I think the cost is not the issue here, as acrylic isn’t that expensive.
but is just not the only problem, every method exposed in the forums has a heavy
trial and error method to follow in order to make it work: FTIR has ir leaking, compliant surface issues, etc.
DI has hotspots, interfence from external IR sources, etc.
At the end, there is no “ultimate/easy” method, since you must make ajudsments for nearly every case.
I think, the less, the better.
With this approach you only track lines (well, sounds simpler than it may be).
Even it could lead to cofigurable resolution (adjusting distance between the deformable surface and a solid panel),
presure sensitive recognition, etc.
I’m maybe naive saying this, but detecting lines is way simpler than many other aproaches, and already exists many processing tools for that.
And for an initial start you dont need to know how much the grid deformed, but where (no pressure sensing).
Best regards