This is a tutorial on the algorithm, 'SALSA', to solve L1-norm penalized least square problems. This problem arises in numerous sparsity-based signal processing methods, including denoising, deconvolution, missing data estimation, and signal separation. This tutorial focuses on four formulations:
Download the tutorial: SALSA.pdf (pdf file)
This tutorial is available as a Connexions module
Download the Matlab software: SALSA_toolbox.zip (zip file)
stylesIllustrative examples with real and complex data.
Ivan Selesnick
NYU Polytechnic School of Engineering
Electrical and Computer Engineering Dept.
Brooklyn, New York, USA