I have been researching ways to bring multi-touch input to display walls for the past couple of years. My approach has been to use commodity web cameras to triangulate objects in front of a wall-sized display. The advantage to this approach is that it is easily scalable, and can work with display walls where it is not convenient to actually touch the display. I have written about how the my interface can be used to interact with wall-sized displays before, but now I have also put up a more in-depth technical look at how the system works. You can read about it here.

Here are a couple of images illustrating the camera-setup, and how the cameras can be used to interact with the display wall by triangulating the position of objects in a plane parallel to the display wall's canvas:

16 cameras along the floor of a display wall. The display wall is located at the Department of Computer Science, University of Tromsø, Norway.

Using the cameras to triangulate the position of objects, as well as disambiguating false positives from real objects.