-
Works Quickly
Leverages GAMP's approximate message passing framework for extreme computational efficiency.
-
Highly Flexible
Object-oriented software leads to a modular environment for constructing a variety of structure and observation models.
-
Learns Structure
Expectation-maximization algorithms allow for data-driven learning of model parameters.
Welcome
Welcome to the homepage for the turboGAMP algorithm, a Bayesian compressed sensing (CS) algorithm that is designed to solve structured sparse signal recovery problems both quickly and accurately.
Here you can learn more about the algorithm, find documentation on the theory and practice of applying turboGAMP, and download the MATLAB source code to try it for yourself (and perhaps contribute to the growing code base!).
Latest News
- June 11, 2013
EMturboGAMP v0.2 MATLAB source code is now available for download.
- October 17, 2012
Many additional signal priors and structure models were just added to the Sourceforge subversion repository. Check it out!