Spectral Simplification of Graphs

Abstract

Although inexact graph-matching is a problem of potentially exponential complexity, the problem may be simplified by decomposing the graphs to be matched into smaller subgraphs. If this is done, then the process may cast into a hierarchical framework and hence rendered suitable for parallel computation. In this paper we describe a spectral method which can be used to partition graphs into non-overlapping subgraphs. In particular, we demonstrate how the Fiedler-vector of the Laplacian matrix can be used to decompose graphs into non-overlapping neighbourhoods that can be used for the purposes of both matching and clustering.

Cite

Text

Qiu and Hancock. "Spectral Simplification of Graphs." European Conference on Computer Vision, 2004. doi:10.1007/978-3-540-24673-2_10

Markdown

[Qiu and Hancock. "Spectral Simplification of Graphs." European Conference on Computer Vision, 2004.](https://mlanthology.org/eccv/2004/qiu2004eccv-spectral/) doi:10.1007/978-3-540-24673-2_10

BibTeX

@inproceedings{qiu2004eccv-spectral,
  title     = {{Spectral Simplification of Graphs}},
  author    = {Qiu, Huaijun and Hancock, Edwin R.},
  booktitle = {European Conference on Computer Vision},
  year      = {2004},
  pages     = {114-126},
  doi       = {10.1007/978-3-540-24673-2_10},
  url       = {https://mlanthology.org/eccv/2004/qiu2004eccv-spectral/}
}