Sparse Deconvolution (An MM Algorithm)

This tutorial describes a computationally efficient method for deconvolution of 1-D signals. It is assumed that the convolution system is described by a difference equation. The algorithm is derived using the majorization-minimization optimization approach and exploits the computational efficiency of fast solvers for banded systems of linear equations.

Download tutorial: SparseDeconv.pdf (pdf file)

Download Matlab software: SparseDeconv_software.zip (zip file)

Sparse deconvolution in MATLAB: Example1 (link) Example2 (link)

This tutorial is also a CNX module (link)

The animation shows the sparse deconvolution as a function of lambda.

Ivan W. Selesnick
Polytechnic Institute of New York University
Electrical and Computer Engineering
Brooklyn, New York, United States