GRASP Special Seminar: Svetlana Lazebnik, University Of Illinois At Urbana-Champaign, "Towards Open-Universe Image Parsing With Broad Coverage"

Presenter: Svetlana Lazebnik (Homepage)

Event Dates:
  Monday November 4, 2013 from 1:00pm to 2:30pm

* Alternate Location: Levine 307 (3330 Walnut Street)*

Joint work with J. Tighe

I will present our work on image parsing, or labeling each pixel in an image with its semantic category (e.g., sky, ground, tree, person, etc.). Our aim is to achieve broad coverage across hundreds of object categories in large-scale datasets that can continuously evolve. I will first describe our baseline nonparametric region-based parsing system that can easily scale to datasets with tens of thousands of images and hundreds of labels. Next, I will describe our approach to combining this region-based system with per-exemplar sliding window detectors to improve parsing performance on small object classes, which achieves state-of-the-art results on several challenging datasets. Time and strength remaining, I may mention new extensions just submitted to CVPR.

Presenter's Biography:

Svetlana Lazebnik received her Ph.D. at UIUC in May 2006 under the supervision of Prof. Jean Ponce. From August 2007 to December 2011, she was an assistant professor at the University of North Carolina at Chapel Hill, and as of January 2012, has returned as faculty to UIUC. Lazebnik's research specialty is computer vision. The main themes of her research include object recognition and scene interpretation, modeling and organization of large-scale photo collections, and machine learning techniques for visual recognition problems.

Lazebnik's research is supported by the National Science Foundation under grants IIS 1228082 (project webpage) and IIS 0916829, Microsoft Research, Xerox, ARO, the DARPA Computer Science Study Group, and the Sloan Foundation.