Games richer, more dynamic, and more efficient to develop. His work on simulations of crowds and human motions could help game developers—and eventually even players themselves—create believable and compelling virtual characters. Popovic’s work in fluid dynamics is used by ESPN to cover NASCAR, and other pieces of his work will appear in future games from Electronic Arts.
—Brian Curless, an expert in graphics and computer vision, discussed his work on human-body and face-shape modeling and computational photography. He and his colleagues have developed a scanner that captures high-resolution models of human faces at video rates, enabling 3-D facial expression modeling and animation. His other projects combine high-resolution photographs with low-resolution videos to enable things like high-resolution video reconstruction, or editing videos as easily as you would edit a photograph (removing an object, say).
—Steve Seitz, a graphics and imaging guru who helped develop the technology behind Microsoft’s Photosynth, showed how to reconstruct a 3-D model of Rome (as well as smaller, everyday objects) using photos downloaded from Flickr. He and Curless also demonstrated how future gamers may be able to project themselves and their environments into virtual worlds, using camera phones to capture the necessary visual data about themselves and their surroundings.
In my time, I’ve seen dozens of really cool academic projects in graphics, imaging, and robotics that would never see the light of day in a commercial game, despite their best intentions. Because of the caliber of the UW labs and the nature of their applications, I wonder if they’ve actually turned the corner on some very tough problems that will have major implications for game developers. As is usually the case with these projects, their commercial success will come down to cost, ease of use, and how good the results look on screen.
