Online Algorithm for Bayesian Compressed Sensing Reconstruction by Paulo V. Rossi (University of Sao Paulo) and Yoshiyuki Kabashima (Tokyo Institute of Technology)
http://journals.aps.org/pre/abstract/10.1103/PhysRevE.94.022137
This Python code runs the algorithm described in 'Bayesian Online Compressed Sensing' (DOI: https://doi.org/10.1103/PhysRevE.94.022137). It provides an efficient and accurate way to reconstruct a sparse signal in a bayesian online fashion.
To use it, install the required packages at requirements.txt through
pip install -r requirements.txt
and run the script with
python online_CS.py
All simulation parameters can be set in the script itself inside the functions simulation and prior.
If everything goes smooth, it will plot the reconstructed vector and an estimate of the error.