Middlebury MRF Energy Minimization Page

Rick Szeliski (Microsoft Research), Ramin Zabih (Cornell University), Daniel Scharstein (Middlebury College),
Olga Veksler (University of Western Ontario), Vladimir Kolmogorov (University College London),
Aseem Agarwala (University of Washington), Marshall Tappen (MIT), and Carsten Rother (Microsoft Research).

This site contains the code and results (images and plots) accompanying our paper

A Comparative Study of Energy Minimization Methods for Markov Random Fields
R. Szeliski, R. Zabih, D. Scharstein, O. Veksler, V. Kolmogorov, A. Agarwala, M. Tappen, and C. Rother.
In Ninth European Conference on Computer Vision (ECCV 2006), volume 2, pages 19-26, Graz, Austria, May 2006.


The MRF energy minimization library: MRF1.6.zip. View the README file for instructions on usage and proper citation.

Note: The current version contains ICM, Graph cuts swap and expansion moves, and max-product belief propagation. It does not yet contain TRW / TRW-S, which we hope to add in a future version. We will also add more sample applications that use the library.


Stereo matching problems:

Photomontage problems:

Binary image segmentation problems:

Image restoration and inpainting problem:

Support for this work was provided in part by NSF grant IIS-0413169. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.