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.
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).
The implementation of the GED algorithm, is provided by the ged package. A demo is provided in the iPython notebook demo.ipynb.
This software uses Python 2.7 and depends on the following Python packages:
- scikit-learn
- numpy
- networkx
- nibabel
- nilearn