/dsstar

Demo code for our CVPR 2018 paper: "DS*: Tighter Lifting-Free Convex Relaxations for Quadratic Matching Problems"

Primary LanguageMATLABGNU Affero General Public License v3.0AGPL-3.0

DS*

Simple demo code for the paper

DS*: Tighter Lifting-Free Convex Relaxations for Quadratic Matching Problems, F. Bernard, C. Theobalt, M. Moeller, IEEE Conference on Computer Vision and Pattern Recognition (CVPR). June, 2018

If you use this code you are required to cite this paper.

The code provides the implementation of a convex relaxation of the quadratic assignment problem (QAP), which reads min_X X(:)'WX(:) s.t. X is a permutation matrix

Here, X is an n-by-n permutation matrix, and W is an n^2-by-n^2 matrix that defines the QAP. For further details see the above paper.