Physics Embedded Machine Learning for Electromagnetic Data Imaging

Abstract

Electromagnetic (EM) imaging is widely applied in sensing for security, biomedicine, geophysics, and various industries. It is an ill-posed inverse problem whose solution is usually computationally expensive.Machine learning (ML) techniques and especially deep learning (DL) show potential in fast and accurate imaging. However, the high performance of purely data-driven approaches relies on constructing a training set that is statistically consistent with practical scenarios, which is often not possible in EM imaging tasks. Consequently, generalizability becomes a major concern. On the other hand, physical principles underlie EM phenomena and provide baselines for current imaging techniques. To benefit from prior knowledge in big data and the theoretical constraint of physical laws, physics embedded ML methods for EM imaging have become the focus of a large body of recent work. This article surveys various schemes to incorporate physics in learning-based EM imaging. We first introduce background on EM imaging and basic formulations of the inverse problem. We then focus on three types of strategies combining physics and ML for linear and nonlinear imaging and discuss their advantages and limitations. Finally, we conclude with open challenges and possible ways forward in this fast-developing field. Our aim is to facilitate the study of intelligent EM imaging methods that will be efficient, interpretable and controllable.

Electromagnetic imaging and deep learning for transition to renewable energies: a technology review

“…This approach can fully use DL advantages and provide numerical approximations that meet EM physics. Recently, Guo et al (2022) classified and summarized the coupled physics-DL strategies in EM imaging. Three types of physics-embedded models for EM imaging have been identified:…”

“…3) Learning with physics models: This aims to surrogate models by directly solving the EM responses (Shahriari et al, 2020a). This type is the most complex since it resolves the entire EM physical phenomena (Guo et al, 2022).…”

“…The combination of these technical advancements and the availability of diverse data has created a conducive environment for deploying DL models in EM imaging. DL models have shown remarkable accuracy in numerous published papers, with significant improvements over traditional approaches (Guo et al, 2019;Guo et al, 2019;Guo R. et al, 2020;Guo et al, 2020;Guo et al, 2021a;Guo et al, 2021b;Guo et al, 2022). However, assessing the accuracy of DL models in EM imaging requires careful consideration of various factors.…”

“…DL methods can learn both prior parameter relationships and structural similarity of the multiphysics data sets. Therefore, the accuracy of EM imaging may be greatly improved (Chen et al, 2022;Guo et al, 2022), and the uncertainty of imaging could be reduced (Um et al, 2022). These aspects are crucial for DL decision-making inversion and need further study (Oh and Byun, 2021).…”

Spectrogram Inversion for Reconstruction of Electric Currents at Industrial Frequencies: A Deep Learning Approach

“…Machine learning comes in handy for this situation, by using the waterfall images to treat and transfer them to RGB matrices. RGB matrices give definition to each pixel in the image [ 13 ]. The latter will be used for training the CNN that has, as an input, the RGB matrices and the output is regression followed by the estimation of the current (either amplitude or frequency).…”