Alexei “alyosha” Efros

Carnegie Mellon University

Abstract

Reasoning about a scene from a photograph is an inherently ambiguous task. This is because a single image in itself does not carry enough information to disambiguate the world that it is depicting. Of course, people have no problems understanding photographs because of all the prior visual experience they can bring to bear on the task. Moreover, there is strong evidence that human visual perception is largely qualitative in nature — less a measuring device than a mechanism for offering “hints” to the internal visual representation.

 

For computer vision, the questions of representation and the right use of prior visual data seem to me to be the paramount if we are to approach human-level image understanding. In this talk, I will present a high-level overview of our efforts in qualitative scene representation and data-driven approaches for addressing a number of difficult image understanding tasks. Results will be presented for inferring geometric, photometric, semantic and geographic scene properties from a single image.

 

The talk will cover joint work with Derek Hoiem, Abhinav Gupta, Martial Hebert, James Hays, Jean-Francois Lalonde, and Srinivas Narasimhan.

 

Bio

Alexei “Alyosha” Efros is a Finmeccanica associate professor at the Robotics Institute and the Computer Science Department at Carnegie Mellon University. His research is in the area of computer vision and computer graphics, especially at the intersection of the two. He is particularly interested in using data-driving techniques to tackle problems which are very hard to model parametrically but where large quantities of data are readily  available. Alyosha received his PhD in 2003 from UC Berkeley under Jitendra Malik, and spent the following year as a post-doctoral fellow in Andrew Zisserman’s group in Oxford, England. Alyosha is a recipient of CVPR Best Paper Award (2006), NSF CAREER award (2006), Sloan Fellowship Honorable Mention (2010), an SIGGRAPH Significant New Researcher Award (2010).

4:15pm

B17 Upson Hall

Thursday, September 23, 2010

Refreshments at 3:45pm in the Upson 4th Floor Atrium

 

Computer Science

Colloquium

Fall 2010

www.cs.cornell.edu/events/colloquium

Qualitative Data-driven image understanding