README.md

HRP

Code for paper "Efficient Sparse Coding using Hierarchical Riemannian Pursuit," in IEEE Transactions on Signal Processing, Y. Xue, V. K. N. Lau and S. Cai, doi: 10.1109/TSP.2021.3093769.[paper]

If you find they are useful, please cite:

@ARTICLE{9470930,
  author={Xue, Ye and Lau, Vincent K. N. and Cai, Songfu},
  journal={IEEE Transactions on Signal Processing}, 
  title={Efficient Sparse Coding using Hierarchical Riemannian Pursuit}, 
  year={2021},
  volume={},
  number={},
  pages={1-1},
  doi={10.1109/TSP.2021.3093769}}

1.0 Prerequisites

• Matlab

• KSVD Matlab toolbox （for Baseline 1)

Download KSVD v13 from https://www.cs.technion.ac.il/~ronrubin/software.html and install (OMP-Box v10 is required).

• SPAMS Matlab toolbox v2.6 （for Baseline 2)

Download SPAMS from http://spams-devel.gforge.inria.fr/downloads.html. Follow the steps in https://github.com/xhm1014/spams-matlab-install-on-win10 to install.

• CVX Matlab toolbox （for Baseline 4)

Download CVX toobox from http://cvxr.com/cvx/ and install.

2.0 Generate the results for the convergence curves

Run Converge_sim.m in the folder curve_convergence.

3.0 Generate the results for the sample complexity curves

Run Sample_sim.m in the folder curve_samplecomplexity.

4.0 Generate the results for the RMSE heatmap with synthetic data

• Unzip the .zip files in the folder heatmap_synthetic.
• Run Syndata_main.m in the folder heatmap_synthetic.

5.0 Generate the results for the table with real-world sensor data

• Unzip all the .zip files in the folder table_sensor.
• Run Sensor_Data_main.m in the folder table_sensor.
• Raw data of the sensor readings of the Airly network can be downloaded from https://www.kaggle.com/datascienceairly/air-quality-data-from-extensive-network-of-sensors.