International Conference on Scale Space and Variational Methods in Computer Vision

SSVM 2015: Scale Space and Variational Methods in Computer Vision pp 231-242

Probabilistic Correlation Clustering and Image Partitioning Using Perturbed Multicuts

  • Jörg Hendrik Kappes
  • Paul Swoboda
  • Bogdan Savchynskyy
  • Tamir Hazan
  • Christoph Schnörr
Conference paper

DOI: 10.1007/978-3-319-18461-6_19

Volume 9087 of the book series Lecture Notes in Computer Science (LNCS)
Cite this paper as:
Kappes J.H., Swoboda P., Savchynskyy B., Hazan T., Schnörr C. (2015) Probabilistic Correlation Clustering and Image Partitioning Using Perturbed Multicuts. In: Aujol JF., Nikolova M., Papadakis N. (eds) Scale Space and Variational Methods in Computer Vision. SSVM 2015. Lecture Notes in Computer Science, vol 9087. Springer, Cham

Abstract

We exploit recent progress on globally optimal MAP inference by integer programming and perturbation-based approximations of the log-partition function. This enables to locally represent uncertainty of image partitions by approximate marginal distributions in a mathematically substantiated way, and to rectify local data term cues so as to close contours and to obtain valid partitions. Our approach works for any graphically represented problem instance of correlation clustering, which is demonstrated by an additional social network example.

Keywords

Correlation clustering Multicut Perturb and MAP 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Copyright information

© Springer International Publishing Switzerland 2015

Authors and Affiliations

  • Jörg Hendrik Kappes
    • 1
  • Paul Swoboda
    • 2
  • Bogdan Savchynskyy
    • 1
  • Tamir Hazan
    • 3
  • Christoph Schnörr
    • 1
    • 2
  1. 1.Heidelberg Collaboratory for Image ProcessingHeidelberg UniversityHeidelbergGermany
  2. 2.Image and Pattern Analysis GroupHeidelberg UniversityHeidelbergGermany
  3. 3.Department of Computer ScienceUniversity of HaifaHaifaIsrael