Logo: University of Southern California

Computer Science Chair Wins "Most Influential Paper of Decade" Award

Honored for machine vision work teaching computers to recognize and create images of objects
Eric Mankin
February 20, 2007 —
Machine Visonary: Medioni
Gérard Medioni, chair of the Viterbi School department of computer science, has won a “Most Influential Paper of the Decade” award from peers in his specialty for a classic paper he presented eleven years ago.

Medioni has been invited to Tokyo to receive his award in person at the Machine Vision Applications international conference May 16-18. Medioni presented the paper for which he is being honored, “Using Computer Vision in Real Applications: Two Success Stories” at the MVA conference in 1996.

A decade ago, Medioni was engaged in a highly productive line of research in computer vision that led, among other applications, to the techniques now used to integrate images into live video, so that viewers of a football game can see the line that must be crossed to get a first down.

The achievements grew out of algorithms Medioni created enabling computers to recognize objects in video, first looking for edges, then using those edges to define shapes, and then to recognize those shapes when viewed from different angles.

A breakthrough — now being recognized — came when he began to use the same algorithms to guide computers in creating images instead of interpreting them. In addition to the ability to insert images in live video, other applications included:
  • A "3-D pen" that enables a visual designer to create and mold 3-D shapes in a single step, rather than by the traditional means of using many 2-D renderings. The same device can "trace" existing objects in three dimensions to scan them into the computer.
  • A way to combine a minimal number of 2-D photographic views to create a virtual 3-D object, or even a virtual environment.

Medioni holds eight patents and is a Fellow of the American Association for Artificial Intelligence, the Institute of Electrical and Electronics Engineers, and the International Association for Pattern Recognition, and is a winner of the Okawa Foundation Award for Understanding of Information in Images.