Recent progress in Al enhanced electromagnetic inverse scattering problems
Electromagnetic inverse scattering problems (ISPs) aims to determine the spatial position and physical information of scatterers from measured electromagnetic fields data. It has a wide range of applications in many fields such as medical imaging, geophysical exploration, non-destructive imaging and etc. However, due to the ill-posedness and nonlinearity of the electromagnetic inversion problem, accurate and efficient inversion are highly challenging. Many conventional inverse methods have been developed to solve the ISPs, but they generally suffer from high computational cost and slow convergence, hindering its applications. With the rapid development of deep learning methods, deep neural networks (DNNs) have been increasingly explored for ISPs and demonstrate promising performance.
This seminar will review recent progress in AI-enhanced electromagnetic ISPs and present our ongoing work on improving both imaging efficiency and accuracy through physics-driven, unsupervised learning approaches.
Speaker: Prof. YUNYUN HU
College of Electronic and Information Engineering,
Tongji University,
Shanghai 200070, China
e-mail: yunyunhu@tongji.edu.cn
(956 KB)