Comparison Between Conjugate Gradient and Landweber Based Regularization Approaches in $\boldsymbol{L^{p}}$ Banach Spaces for Microwave Tomography

Innovative methods for the tomographic reconstruction of dielectric targets are incessantly proposed by the research community, and one of the main goals is the continuous improvement of both the regularization capabilities and the computational cost [1]–[3]. In this scenario, inversion techniques developed in the mathematical context of $L^{p}$ Banach spaces appear to be particularly effective in the retrieval of the dielectric properties of the targets under test [4]. Initially, Banach space regularization approaches for the nonlinear microwave tomography have been exploited through inexact-Newton techniques that comprise an iterative Newton-based linearization and a truncated Landweber algorithm, which is usually employed for solving the resulting linear problem in a regularized sense.