Modified complex multitask Bayesian compressive sensing using Laplacian scale mixture prior

Bayesian compressive sensing (BCS) is an important sub-class of sparse signal reconstruction algorithms. In this paper, a modified complex multitask (MCMBCS) algorithm using the Laplacian scale mixture (LSM) prior proposed. The LSM first introduced into BCS framework by exploiting its better characteristic and flexibility than traditional prior. Furthermore, integrating out noise variance analytically, MCMBCS significantly improves recovery performance original CMBCS. More importantly, authors not only present iterative but also develop sub-optimal fast implementation method based on marginal likelihood maximisation, which dramatically reduce computational complexity. Finally, sufficient numerical simulations validate proposed in accuracy effectiveness existing work. It revealed that has great potential complex-valued processing field.