Roman Vershynin | Software

  • Regularized Orthogonal Matching Pursuit: a sparse reconstruction algorithm. Matlab code.

    This is a greedy algorithm that can reconstruct any sparse signal from incomplete set of measurements. The reconstruction is exact for any linear measurements satisfying the Uniform Uncertainty Principle.

    This algorithm was developed and analysed in
    this paper joint with my graduate student Deanna Needell. This basic code was written by Deanna Needell. Systematic version is under construction.

  • Kashin's representations and vector quantization. Under construction. Based on this paper joint with Yu. Lyubarskii.