Fast Graph Cuts Using Shrink-Expand Reparameterization

Abstract

Global optimization of MRF energy using graph cuts is widely used in computer vision. As the images are getting larger, faster graph cuts are needed without sacrificing optimality. Initializing or reparameterizing a graph using resultsof a similar one hasprovided efficiencyin the past. In this paper, we present a method to speedup graph cuts using shrink-expandreparameterization. Our scheme merges thenodesof agivengraphto shrink it. The resultinggraph anditsmincutareexpandedandusedtoreparameterizethe originalgraphforfasterconvergence. Graphshrinkingcan be done in different ways. We use a block-wise shrinking similarto multiresolutionprocessingof imagesin ourMultiresolution Cuts algorithm. We also develop a hybrid approach that can mix nodes from different levels without affecting optimality. Our algorithm is particularly suited for processing large images. The processing time on the full detail graph reduces nearly by a factor of 4. The overall application time including all book-keeping is faster by a factorof2onvarioustypesof images.

Cite

Text

Sakurikar and Narayanan. "Fast Graph Cuts Using Shrink-Expand Reparameterization." IEEE/CVF Winter Conference on Applications of Computer Vision, 2012. doi:10.1109/WACV.2012.6163025

Markdown

[Sakurikar and Narayanan. "Fast Graph Cuts Using Shrink-Expand Reparameterization." IEEE/CVF Winter Conference on Applications of Computer Vision, 2012.](https://mlanthology.org/wacv/2012/sakurikar2012wacv-fast/) doi:10.1109/WACV.2012.6163025

BibTeX

@inproceedings{sakurikar2012wacv-fast,
  title     = {{Fast Graph Cuts Using Shrink-Expand Reparameterization}},
  author    = {Sakurikar, Parikshit and Narayanan, P. J.},
  booktitle = {IEEE/CVF Winter Conference on Applications of Computer Vision},
  year      = {2012},
  pages     = {65-71},
  doi       = {10.1109/WACV.2012.6163025},
  url       = {https://mlanthology.org/wacv/2012/sakurikar2012wacv-fast/}
}