Most of the research in compressed sensing type applications now is focused on more sophisticated prior distributions than sparse/Laplacian. These more general bayesian approaches provide significant performance improvements over LASSO etc.