Chapter 21: Plug-and-Play and Regularization by Denoising

Learning Objectives

• Derive PnP-ADMM and PnP-PGD by replacing the proximal operator with a denoiser
• Explain how deep denoisers (DnCNN, DRUNet, SwinIR) serve as implicit image priors
• State convergence conditions for PnP with non-expansive and gradient-step denoisers
• Define the RED objective and compute its gradient under Jacobian symmetry
• Apply PnP and RED to RF imaging problems and compare with LASSO and OAMP