Frustrated Total Internal Reflection, Diffuse Illumination, and Capacitive or Resistive Sensing
There have been quite a few choices when it comes to MT, or multi-touch sensing over the years, as chronicled by Bill Buxton. All of the different methods each have their own pros and cons for different environments and applications.
The FTIR, or Frustrated Total Internal Reflection was popularized by Jeff Han in his 2006 TED presentation.
Image credit Jeff Han http://www.cs.nyu.edu/~jhan/ftirsense/ftirschematic.gif
This image has been found to be an incomplete representation of a working FTIR MT screen. Most FTIR screens include a compliant surface (Sorta-Clear 40, Transparent Silicone Caulk, or Lexel) directly applied to the front surface of the acrylic and covered with a projection surface (Rosco Grey, trace paper, or Mylar). If you would like to see how the FTIR principle works, then click here for a Java applet demo.
Diffuse Illumination works two ways. The first is by filling a closed box with "ambient" or "diffuse" IR light which provides a uniform IR brightness evenly across the surface of the acrylic and is reflected back strongly when someone touches it, known as rear diffuse illumination (RDI). The other method involves the exact opposite IR light source. Front diffuse illumination (FDI) uses IR illumination from outside the box, typically sunlight, and it is relatively trivial to switch back and forth between a front and rear illumination setup without adjusting any hardware. This may be why it is currently being used by Citywall, an outdoor public MT exhibit.
A diffusing layer (aka standard art store trace paper) is applied to the bottom of the acrylic, just like the image above, and it also acts as a projection surface. This is similar to the FTIR method, except the IR light source comes from inside (or outside) of the box instead of the edge of the acrylic. With so much of the IR light escaping through the front of the acrylic, unless reflected by an object on the surface, I worry about long exposure to strong IR light damaging people's vision. While this method enables easy fiducial marker tracking, it can be difficult to properly calibrate all the components with the current software available or eliminate "noise" from anything too reflective above (but not touching) the surface.
Using a RDI setup, reacTable tracks multiple fiducial markers and fingers at http://reactable.iua.upf.edu/
DI can be the cheapest and easiest to initially setup and get working, it requires the least time and materials, but it is more suited to rigid, self-contained, permanent installations with low ambient IR light levels and no direct sunlight (for the most common RDI setups only).
Capacitive or Resistive Sensors use expensive, proprietary hardware and software to calculate a finger press based on triangulating "pressure waves" traveling through the acrylic or glass. The most popular for large displays are from 3M and they are just no fun. It's like the corporate focus group version of fun.
Image credit 3M
Since there is no camera, there is no fiducial marker tracking, although you could add a camera just for this purpose, you would have to question your sanity by increasing the necessary processing power to deal with the glut of information, in addition to the (amazingly slow) latency of the Capacitive or Resistive sensors as they struggle with more than two points of contact (if they can even handle more; the iPhone can't). I have not built a MT display with these sensors, but I have used one, and if I'm not mistaken, it's what the iPhone and most consumer POS touch screens currently use. Yawn. Other than that, you can investigate yourself online, or catch up with Bill Buxton's Multi-Touch research.
