Tech Student Wins Prestigious Fellowship to Research Computer Vision
First Time a Computer Science or Engineering Student Selected
Posted January 21, 2004 | Atlanta
Somewhere between scaling sheer cliffs, traveling the world and swing dancing the night away, Georgia Tech computer science Ph.D. candidate, Gabriel J. Brostow, learned enough about computer vision not only to get a Ph.D. from Georgia Tech's College of Computing, but he also won one of only two Marshall Sherfield Fellowships awarded to American science and engineering students to study at a British university. He is the first computer science or engineering student selected since the award was created in 1998.
"The opportunity to work with another group of leading scientists in my field at the conclusion of my Ph.D. will be a welcome time of productivity and adjustment of my own long-term research objectives, before I return to the U.S. seeking a faculty position," said Brostow.
The Marshall Aid Commemoration Commission offers up to two post-doctoral Fellowships in science and engineering, in addition to the forty well-known Marshall Scholarships for undergraduates to pursue graduate studies. The aim of the Marshall Sherfield Fellowships is to introduce American scientists and engineers to the cutting edge of U.K. science and engineering. The intention is to build longer-term contacts and international links between the U.K. and the United States in key scientific areas.
Brostow, on track to complete his Ph.D. in May, plans to spend the year at the University of Cambridge working with Professor Robert Cipolla in the Vision and Robotics group. He would like to stay for a second year and join one of the Cambridge colleges as well as pursue further collaborations between computer vision and other departments at Cambridge and elsewhere in the U.K.
Computer vision initially started as a sub-domain of artificial intelligence but has grown to include numerous researchers focusing on medical imaging, machine learning, forensics, robotics, and graphics to name a few. The field borrows from and contributes to these core areas and is rapidly expanding due to the use of computer-controlled video cameras gaining acceptance in many fields as a primary data source, says Brostow. More specifically, Brostow's own research has broad appeal since it can be employed in security and biomechanical applications as well as other areas.
Brostow completed his B.S. in Electrical Engineering from the University of Texas at Austin and worked at Lockheed in New Hampshire, before coming to Georgia Tech in 1997 to pursue an M.S. in Human Computer Interaction. His thesis adviser, Associate Professor Irfan Essa, convinced him to stay for a Ph.D. Brostow works closely with Essa in the area of computer vision and animation. Over the last six years he assisted Essa in creating and then eventually co-taught the very popular CS course on Digital Video Special Effects.
"In the last few years Gabe has proven himself to be an able researcher, mastering many technical areas, and a great motivator, working with many students on a variety of projects," said Essa. "He has been involved in several research projects with me, where he has taken the lead and also recruited others to work with him. When I saw the announcement of the fellowship, which asked for ambassadors to the U.K. to represent the U.S. and their research field, I had no hesitation in recommending him as he fits the bill entirely."
Brostow has held summer research positions with IBM, Microsoft Research, and in the Research & Development group at Industrial Light & Magic. His long-term plans are to return to academia.
Brostow has a very international outlook having lived in Canada, Mexico and Germany. In addition, his parents are both from Poland, and he speaks both Polish and German. He has traveled extensively to most of Europe, several former Soviet Republics, Iceland, Brazil, South Africa, Israel, India, Thailand, South Korea, Japan, Australia and New Zealand. Starting in October, he'll be able to add the United Kingdom to his list of homes.