/brain_ged

Graph matching of brain connectivity networks based on graph edit distance

Primary LanguageJupyter NotebookMIT LicenseMIT

Finding correspondences between brain connectivity networks using graph edit distance

This code provides a Python implementation of constrained graph edit distance using the Hungarian algorithm (Riesen and Bunke, 2009) to identify correspondences between brain connectivity networks and obtain an estimate of their (dis)similarity.

overview

Please cite the corresponding paper when using the code:

Ktena, S.I., Arslan, S., Parisot, S., Rueckert, D.: Exploring Heritability of Functional Brain Networks with Inexact Graph Matching. In: IEEE International Symposium on Biomedical Imaging (ISBI) (2017).

correspondences

Usage

The implementation of the GED algorithm, is provided by the ged package. A demo is provided in the iPython notebook demo.ipynb.

Dependencies

This software uses Python 2.7 and depends on the following Python packages:

  • scikit-learn
  • numpy
  • networkx
  • nibabel
  • nilearn