C-SBL(VB): A Variational Bayes Algorithm for Sparse Recovery of Signals with Unknown Clustering Pattern

We develop a new algorithm to solve the inverse problem of compressive sensing for sparse signals with unknown clustering pattern in the presence of noise. The proposed algorithm is based on Bernoulli-Gaussian prior modeling and variational Bayes inference. The proposed algorithm differs from other works in this area due to the way the clustering pattern is learned, in that we account for the unknown clustering pattern via one prior defined on the supports of the solution rather than having three sets of hyperpriors on either the supports or the precisions of the solution.