Abstract: We propose an sparse Bayesian learning (SBL)based method that leverages group sparsity and multiple parameterized dictionaries to detect the relevant dictionary entries and estimate their ...