ML Anthology

Generalized Roof Duality for Multi-Label Optimization: Optimal Lower Bounds and Persistency

Thomas Windheuser, Hiroshi Ishikawa, Daniel Cremers
ECCV 2012 pp. 400-413
doi:10.1007/978-3-642-33783-3_29 /eccv/2012/windheuser2012eccv-generalized/

Abstract

We extend the concept of generalized roof duality from pseudo-boolean functions to real-valued functions over multi-label variables. In particular, we prove that an analogue of the persistency property holds for energies of any order with any number of linearly ordered labels. Moreover, we show how the optimal submodular relaxation can be constructed in the first-order case.

PDF ECCV Semantic Scholar

Cite

Text

Windheuser et al. "Generalized Roof Duality for Multi-Label Optimization: Optimal Lower Bounds and Persistency." European Conference on Computer Vision, 2012. doi:10.1007/978-3-642-33783-3_29

Markdown

[Windheuser et al. "Generalized Roof Duality for Multi-Label Optimization: Optimal Lower Bounds and Persistency." European Conference on Computer Vision, 2012.](https://mlanthology.org/eccv/2012/windheuser2012eccv-generalized/) doi:10.1007/978-3-642-33783-3_29

BibTeX

@inproceedings{windheuser2012eccv-generalized,
  title     = {{Generalized Roof Duality for Multi-Label Optimization: Optimal Lower Bounds and Persistency}},
  author    = {Windheuser, Thomas and Ishikawa, Hiroshi and Cremers, Daniel},
  booktitle = {European Conference on Computer Vision},
  year      = {2012},
  pages     = {400-413},
  doi       = {10.1007/978-3-642-33783-3_29},
  url       = {https://mlanthology.org/eccv/2012/windheuser2012eccv-generalized/}
}