Header logo is

Global Connectivity Potentials for Random Field Models

2009

Conference Paper

ei


Markov random field (MRF, CRF) models are popular in computer vision. However, in order to be computationally tractable they are limited to incorporate only local interactions and cannot model global properties, such as connectedness, which is a potentially useful high-level prior for object segmentation. In this work, we overcome this limitation by deriving a potential function that enforces the output labeling to be connected and that can naturally be used in the framework of recent MAP-MRF LP relaxations. Using techniques from polyhedral combinatorics, we show that a provably tight approximation to the MAP solution of the resulting MRF can still be found efficiently by solving a sequence of max-flow problems. The efficiency of the inference procedure also allows us to learn the parameters of a MRF with global connectivity potentials by means of a cutting plane algorithm. We experimentally evaluate our algorithm on both synthetic data and on the challenging segmentation task of the PASCAL VOC 2008 data set. We show that in both cases the addition of a connectedness prior significantly reduces the segmentation error.

Author(s): Nowozin, S. and Lampert, CH.
Book Title: CVPR 2009
Journal: Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2009)
Pages: 818-825
Year: 2009
Month: June
Day: 0
Publisher: IEEE Service Center

Department(s): Empirical Inference
Research Project(s): Machine Learning for Visual Scene Understanding
Bibtex Type: Conference Paper (inproceedings)

DOI: 10.1109/CVPRW.2009.5206567
Event Name: IEEE Computer Society Conference on Computer Vision and Pattern Recognition
Event Place: Miami Beach, FL, USA

Address: Piscataway, NJ, USA
Digital: 0
Language: en
Organization: Max-Planck-Gesellschaft
School: Biologische Kybernetik

Links: PDF
PDF
Web

BibTex

@inproceedings{5828,
  title = {Global Connectivity Potentials for Random Field Models},
  author = {Nowozin, S. and Lampert, CH.},
  journal = {Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2009)},
  booktitle = {CVPR 2009},
  pages = {818-825},
  publisher = {IEEE Service Center},
  organization = {Max-Planck-Gesellschaft},
  school = {Biologische Kybernetik},
  address = {Piscataway, NJ, USA},
  month = jun,
  year = {2009},
  month_numeric = {6}
}