An interactive art and technology installation
by Trevor Darrell, Gaile Gordon, Mike Harville,
John Woodfill, Harlyn Baker, and Aaron Hertzmann
Fast automated detection and tracking of faces is the technical key to this piece. Using a half-silvered mirror, two cameras observe the viewers along the same optical path that the viewers observe the display. These images are analyzed to track the viewers. Three different techniques are combined to produce robust results: a stereo vision module separates nearby people from the background, a color module detects regions of skin hue, and a pattern module recognizes contrast patterns that appear to be faces. The dynamic graphical effects are created by applying distortions to the live video signal in the regions of the image containing faces. These distortions follow the viewers' faces as they move in front of the display, encouraging the viewer to attempt to manipulate their image through their movements and expressions.
This demonstration shows how computers can begin to observe and analyze visual information quickly enough for use in interactive systems, allowing the user to control a display directly without a mouse or keyboard.
The Tech Museum of Innovation
San Jose, California, USA, 31 October 1998 - 31 March 1999
SIGGRAPH '98 Enhanced Realities
Orlando, Florida, USA, 19 - 24 July 1998
Computer Vision and Pattern Recognition '98
Santa Barbara, California, USA, 23 - 25 June 1998
As "Magic Morphin Mirror"
SIGGRAPH '97 Electronic Garden,
Los Angeles, Calif, USA, 3 - 8 August 1997
T. Darrell, G. Gordon. J. Woodfill, M. Harville, "A Virtual Mirror Interface using Real-time Robust Face Tracking", Proceedings of the the Third International Conference on Face and Gesture Recognition, (Nara, Japan) IEEE Computer Society Press, April 1998.
J. Woodfill, B. Von Herzen, "Real-Time Stereo Vision on the PARTS Reconfigurable Computer," Proceedings IEEE Symposium on Field-Programmable Custom Computing Machines, Napa, pp. 242-250, April 1997.
R. Zabih, J. Woodfill, "Non-parametric Local Transforms for Computing Visual Correspondence", Third European Conference on Computer Vision, (Stockholm, Sweden) May 1994.