This video is a bit on the dry side, but watch it all the way through. It's a demonstration of a new 3D facial rendering technology by Volker BLanz and Thomas Vetter of the MPI for Biological Cybernetics in Germany. It does a lot of amazing things, but the most striking is the ability to map a 2D photograph (or even painting) into a realistic, pose-able, 3D model. In the video, stills of Tom Hanks, Audrey Hepburn, and DaVinci's Mona Lisa are used to create very realistic looking 3D models of the heads. They're even able to change the lighting and have added elements to the pictures (such as a hat) cast realistic shadows.
Combining this technology with rapid prototyping could lead to some very cool physical reconstructions of historical figures. I imagine it could also have a lot of potential for forensic reconstruction. I wonder what would happen if they applied this technology to the image on the Shroud of Turin. Very cool.