We measure values of a real valued smooth signal on the specific nodes of on a network using graph signal sampling.
Using these known values, we can predict the signal values on the remaining nodes of a network.
The software in this repository implements the Approximate volume maximization algorithm for graph signal sampling from this paper.
For this tutorial, I assume you are using a UNIX like OS.
This repository uses the following packages. They need to be downloaded in order to run the software in this respository.
Download and uncompress these packages in any directory on your computer. For this tutorial, I will assume that you placed all downloads in ~/Downloads.
Change directory to ~/Downloads.
cd ~/Downloads
Clone this repository from GitHub.
git clone https://github.com/STAC-USC/Graph-signal-sampling-AVM.git
Change directory to that of the downloaded repository.
cd Graph-signal-sampling-AVM
Start matlab.
matlab &
Run the algorithm at the MATLAB command prompt. The demo function takes two arguments - path of the GSP toolbox, and path of the GraphSamplingBox.
demo_sample_and_reconstruct('~/Downloads/gspbox', '~/Downloads/GraphSamplingBox');
If you're able to run the code succcessfully, you should be able to see results similar to these.
You should be able to see the sampled vertices highlighted in red.
You should be able to see the original and the reconstructed signal comparison in a single plot.
Your appreciation motivates me to do rigorous research and develop quality software. If you find this software useful, you can show your appreciation by starring the repository. If you use this software towards your research, cite this paper. Here's the bibtex citation for your convenience.
@article{JAYAWANT2022108436,
title = {Practical graph signal sampling with log-linear size scaling},
journal = {Signal Processing},
volume = {194},
pages = {108436},
year = {2022},
issn = {0165-1684},
doi = {https://doi.org/10.1016/j.sigpro.2021.108436},
url = {https://www.sciencedirect.com/science/article/pii/S0165168421004734},
author = {Ajinkya Jayawant and Antonio Ortega},
keywords = {Graph, Signal, Sampling, D-optimal, Volume, Coherence},
abstract = {Graph signal sampling is the problem of selecting a subset of representative graph vertices whose values can be used to interpolate missing values on the remaining graph vertices. Optimizing the choice of sampling set using concepts from experiment design can help minimize the effect of noise in the input signal. While many existing sampling set selection methods are computationally intensive because they require an eigendecomposition, existing eigendecomposition-free methods are still much slower than random sampling algorithms for large graphs. In this paper, through optimizing sampling sets towards the D-optimal objective from experiment design, we propose a sampling algorithm that has complexity comparable to random sampling algorithms, while reaching accuracy similar to existing eigendecomposition-free methods for a broad range of graph types.}
}
I welcome technical feedback either through github issues or email to the corresponding author.