Title :
Sparse contrast-source inversion using linear-shrinkage-enhanced inexact Newton method
Author :
Desmal, Abdulla ; Bagci, Hakan
Author_Institution :
Div. of Comput., Electr., & Math. Sci. & Eng., King Abdullah Univ. of Sci. & Technol. (KAUST), Thuwal, Saudi Arabia
Abstract :
A contrast-source inversion scheme is proposed for microwave imaging of domains with sparse content. The scheme uses inexact Newton and linear shrinkage methods to account for the nonlinearity and ill-posedness of the electromagnetic inverse scattering problem, respectively. Thresholded shrinkage iterations are accelerated using a preconditioning technique. Additionally, during Newton iterations, the weight of the penalty term is reduced consistently with the quadratic convergence of the Newton method to increase accuracy and efficiency. Numerical results demonstrate the applicability of the proposed method.
Keywords :
Newton method; electromagnetic wave scattering; microwave imaging; Newton iterations; electromagnetic inverse scattering; ill-posedness; inexact Newton methods; linear-shrinkage-enhanced inexact Newton method; microwave imaging; nonlinearity; preconditioning technique; sparse contrast-source inversion; Algorithm design and analysis; Convergence; Electromagnetics; Microwave imaging; Microwave theory and techniques; Newton method;
Conference_Titel :
Antennas and Propagation Society International Symposium (APSURSI), 2014 IEEE
Conference_Location :
Memphis, TN
Print_ISBN :
978-1-4799-3538-3
DOI :
10.1109/APS.2014.6904763