Improved Microwave Imaging with the Extended Born Iterative Method.

This paper proposes an extension to the Born-Iterative-Method (BIM) algorithm with a goal to enhance its working range of imaging permittivity contrasts. State-of-the-art BIM variants are governed by the weak scattering principle which is limited to applications with low contrasts. The proposed Extended BIM (EBIM) uses the localized non-linear approximation during each Born iteration; an additional computational step which is well compensated by very fast convergence, numerical stability, and accommodation of higher contrasts inside the imaging domain. As a quantitative algorithm, we show that the EBIM provides improved reconstruction of the electric fields, imaging contrasts ≈ 3 times greater than traditional BIM variants. As a qualitative implementation, the EBIM is shown to robustly identify anomalies in differential imaging problems, even with trivial initialization conditions. Though complexity of the anatomy results in loss of quantitative information, the ability to address various imaging scenarios with numerical stability in the presence of a small number of antennas (12 in this study) suggest that the EBIM algorithm can be used as a robust method for microwave imaging.