International Workshop on Graphical Models in Computer Vision
7th of September 2014, Zurich, Switzerland
in conjunction with European Conference on Computer Vision
Aims and scope

Probabilistic graphical models (PGMs) are now ubiquitous in a wide variety of computer vision tasks from low-level and high-level vision problems. They are expected to be of fundamental importance with regard to the task of many computer vision applications, such as denoising, stereo reconstruction, object segmentation, scene understanding, and human activity recognition. The purpose of this workshop is to bring together an examination of recent advances in PGMs with emerging problem formulations motivated by Computer Vision applications.

Submissions are invited from all areas of computer vision relevant for graphical models. Topics of interest include, but are not limited to:

  • Modeling aspects in PGMs
  • Inference methods for higher-order models
  • MAP inference with unknown graph structure
  • Inference for large scale PGMs
  • Inference in hybrid continuous-discrete models
  • Learning methods, including partially and weakly labeled data
  • Distributed inference and learning techniques
  • Anytime algorithms for inference and learning

All manuscripts will be subject to a double-blind review process. In the proceedings, accepted papers will be allocated 14 pages for the main paper excluding references.

  • 08:30-09:00 - Registration
  • 09:00-09:10 - Introduction: Sebastian Nowozin (MSRC)
  • 09:10-10:10 - Keynote talk: Vittorio Ferrari (University of Edinburgh)
  • 10:10-10:35 - Varun K. Nagaraja Vlad I. Morariu Larry S. Davis. Feedback Loop between High Level Semantics and Low Level Vision
  • 10:35-11:10 Coffee break
  • 11:10-12:10 - Keynote talk: Raquel Urtasun (University of Toronto)
  • 12:10-14:00 Lunch Break
  • 14:00-15:00 - Keynote talk: Joerg Hendrik Kappes (University of Heidelberg)
  • 15:00-15:25 - Cheng Zhang Hedvig Kjellstroem. How to Supervise Topic Models
  • 15:25-16:00 Coffee Break
  • 16:00-17:00 - Keynote talk: Pushmeet Kohli (MSRC)
  • 17:00-17:25 - Nathan Silberman, David Sontag, Rob Fergus. Instance Segmentation of Indoor Scenes using a Coverage Loss
  • 17:25-17:50 - Joerg Hendrik Kappes, Thorsten Beier and Christoph Schnoerr. MAP-Inference on Large Scale Higher-Order Discrete Graphical Models by Fusion Moves
  • 17:50-18:00 - Closing remarks
Keynote Speakers