ML Anthology

A Convex Relaxation for Multi-Graph Matching

Paul Swoboda, Dagmar Kainm"uller, Ashkan Mokarian, Christian Theobalt, Florian Bernard
CVPR 2019
doi:10.1109/CVPR.2019.01141 /cvpr/2019/swoboda2019cvpr-convex/

Abstract

We present a convex relaxation for the multi-graph matching problem. Our formulation allows for partial pairwise matchings, guarantees cycle consistency, and our objective incorporates both linear and quadratic costs. Moreover, we also present an extension to higher-order costs. In order to solve the convex relaxation we employ a message passing algorithm that optimizes the dual problem. We experimentally compare our algorithm on established benchmark problems from computer vision, as well as on large problems from biological image analysis, the size of which exceed previously investigated multi-graph matching instances.

PDF CVPR Semantic Scholar

Cite

Text

Swoboda et al. "A Convex Relaxation for Multi-Graph Matching." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019. doi:10.1109/CVPR.2019.01141

Markdown

[Swoboda et al. "A Convex Relaxation for Multi-Graph Matching." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019.](https://mlanthology.org/cvpr/2019/swoboda2019cvpr-convex/) doi:10.1109/CVPR.2019.01141

BibTeX

@inproceedings{swoboda2019cvpr-convex,
  title     = {{A Convex Relaxation for Multi-Graph Matching}},
  author    = {Swoboda, Paul and Kainm"uller, Dagmar and Mokarian, Ashkan and Theobalt, Christian and Bernard, Florian},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
  year      = {2019},
  doi       = {10.1109/CVPR.2019.01141},
  url       = {https://mlanthology.org/cvpr/2019/swoboda2019cvpr-convex/}
}