Aaron Bobick

Aaron Bobick
TSRB Office 211B

Aaron Bobick is Professor and former Chair of the School of  Interactive Computing. He has B.Sc. degrees from MIT in Mathematics (1981) and Computer Science (1981) and a Ph.D. from MIT in Cognitive Science (1987). He is a pioneer in the area of action recognition by computer vision, having authored over 80 papers in this area alone. He joined the MIT Media Laboratory faculty in 1992 where he led the Media Lab DARPA VSAM project as well as a Dynamic Scene Analysis effort funded by the CIA.  In 1999 Prof. Bobick joined the faculty at Georgia Tech where he became the Director of the GVU Center, an internationally known research center in computer vision, graphics, ubiquitous computing, and HCI. As Director he greatly expanded the scope of GVU both in the direction of human expression – such as music technology or narrative – and in computers interacting with the outside world both through sensing and robotics. In 2005, the unitary College of Computing was divided into independent departments and the School of Interactive Computing was created with Prof. Bobick serving as the founding Chair. Under his leadership the School launched several novel computing degrees including BS in Computational Media as well as PhD programs in Robotics and in Human Centered Computing. Prof. Bobick has served as a senior area chair for most international computer vision conferences and was Program Chair of IEEE Conference on Computer Vision and Pattern Recognition. He is an ACM Distinguished Scientist and has also served on the advisory board or boards of directors of a variety of surveillance-focused computer vision and medical imaging technology companies.
Prof. Bobick's research spans a variety of aspects of computer vision. His early work focused on human activity recognition from  video sequences: three examples are the basic recognition of human movements, natural gesture understanding, and the classification of football plays. Each of these examples requires describing human activity in a manner appropriate for the domain, and developing recognition techniques suitable for those representations. An intriguing application demonstration of this technology was the the KidsRoom, the world's first vision-aware, interactive narrative play-space for children. The room employed large-scale video and sound to take the children through a fantasy story; all the sensing was accomplished using computer vision. Recently, Prof. Bobick has shifted towards computer vision for robots, in particular robots understanding how they can interact with objects and people. As robots move from isolated factory environments to human spaces they will need to collaborate with humans and manipulate the objects found in the human world. Prof. Bobick’s research in behavior modeling and prediction for collaboration along with the perceptual analysis of object affordances is targeted at addressing this transition in robotics.


GVU Center
Robotics and Intelligent Machines Center
School of Interactive Computing