Compressed Sensing : Balanced Graph Matching (Spring 2019)

As part of Compress Sensing course, you’ll find in this repo our project report. We studied the two following articles: « Balanced Graph Matching » from Cour et al. (2005) and « A Fast Semidefinite Approach to Solving Binary Quadratic Problems » from Wang et al. (2013). Our notebook provides a full theoretical review of Cour et al.’s article along with our implementation. For more convenience, you can use NbViewer to view it: https://nbviewer.jupyter.org/github/afiliot/Compressed_sensing_project/blob/master/report.ipynb?flush_cache=true. We incorporate SDCut algorithm from Wang et al. adapted to one-to-one matching problem. The folder pickle contains useful files for quickly visualising the results from our experiments.