Performance Analysis of Iterative Methods in Microwave Imaging With Various Regularization Techniques

Quantitative microwave imaging (MWI) involves solving the inverse scattering problem (ISP), which is characterized by nonlinearity and ill-posedness. To address the challenges posed by ISP, iterative linearization techniques have been introduced alongside regularization procedures. Born Iterative Method (BIM) and Distorted Born Iterative Method (DBIM) are well-established approaches in the field of microwave imaging. The primary objective of our study was to conduct a comparative analysis to evaluate the performance of traditional regularization techniques such as Truncated Singular Value Decomposition (TSVD), Tikhonov regularization, and the truncated Landweber algorithm with choosing different regularization parameters, within the framework of these established methods. The lack of explicit mention of optimal parameter values and comparison of regularization methods highlights a gap in the existing literature. Investigating and comparing different regularization methods and their corresponding optimal parameter values for different structures and features within the framework of linearization methods can provide valuable insights into effectively solving inverse problems in MWI. Additionally, exploring how different regularization parameters impact the accuracy and stability of the solutions obtained through BIM and DBIM can help researchers and practitioners make informed decisions to choose a regularization method and its corresponding parameter value for a specific problem. Our research aimed to provide a comprehensive baseline that would be beneficial for future studies and practical applications in microwave imaging.