[Colloquium] TTIC Colloquium: Adarsh Kowdle, Cornell

[Liv Leader]

When:     Friday, June 17th @ 12:30 p.m.


Where:   TTIC Conference Room #526, 6045 S. Kenwood Avenue, 5th Floor

Who:      Adarsh Kowdle, Cornell

Title:      Putting humans into the loop for image based modeling



The task of image based modeling or recovering the 3D structure of a scene
via multi-view 2D images is not new to computer vision. While a number of
successful automatic algorithms have been proposed for achieving this, scene
irregularities such as textureless surfaces, specular surfaces, and lack of
geometric cues, hinder their performance. On the other hand, when we humans
look at a scene we can much better discern the geometric structure that
underlies the pixel data we view. Interactive 3D reconstruction algorithms
try to exploit this, by requiring the user to provide interactions in the
form of line drawings, 2D polygons, etc. While these algorithms work well,
they are not only cumbersome for a user, but also seem unnecessary given the
performance of the automatic algorithms.



In this talk, we will first take a brief look at recent automatic image
based modeling approaches, followed by interactive image based modeling
algorithms. In the second half of the talk, I will focus on our recent work
on active learning for piecewise planar 3D reconstruction where, we exploit
the performance of the automatic algorithm and intelligently put the user
into the loop.



http://chenlab.ece.cornell.edu/people/adarsh/



Hosts:  Devi Parikh, dparikh at ttic.edu and Dhruv Batra, dbatra at ttic.edu

-- 
Liv Leader
Faculty Services

Toyota Technological Institute
6045 S Kenwood Ave, #504
Chicago, IL 60637
Phone- (773) 834-2567
Fax-     (773) 834-9881
Email-  lleader at ttic.edu <jam at ttic.edu>
Web-   www.ttic.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20110614/aaaa9162/attachment.htm