Abstract:This letter proposes a novel deep learning (DL) approach to resolve the electromagnetic inverse scattering (EMIS) problems. The conventional approaches of resolving EMIS problems encounter assorted difficulties, such as high contrast, high computational cost, inevitable intrinsic nonlinearity, and strong ill-posedness. To surmount these difficulties, a novel DL approach is proposed based on a novel complex-valued deep fully convolutional neural network structure. The proposed complex-valued deep learning model for solving EMIS problems composes of an encoder network and its corresponding decoder network, followed by a final pixel-wise regression layer. The complex-valued encoder network extracts feature fragments from received scattered field data, while the role of the complex-valued decoder network is mapping the feature fragments to retrieve the final contrasts (permittivities) of dielectric scatterers. Hence, the proposed deep learning model functions as an "heterogeneous" transformation process, where measured scattering field data is converted into the corresponding contrasts of scatterers. As a consequence, the EMIS problem could be resolved accurately even for extremely high-contrast targets. Numerical benchmarks have illustrated the feasibility and accuracy of our proposed approach. The proposed approach opens a novel path for the deep learning-based real-time quantitative microwave imaging for high-contrast scatterers.

Electromagnetic Modeling Using an FDTD-Equivalent Recurrent Convolution Neural Network: Accurate Computing on a Deep Learning Framework.

An End-to-End Neural Network for Complex Electromagnetic Simulations

Study on a Recurrent Convolutional Neural Network Based FDTD Method

A Parallel FDTD Scheme for Electromagnetic Analysis and Design of MRI System.

Modeling of EMI Filter by Means of Recurrent Neural Network with Encoding Layer

Modeling method of finite-difference time-domain based on frequency network models of active devices

Enhanced Deep Learning Approach Based on the Deep Convolutional Encoder–Decoder Architecture for Electromagnetic Inverse Scattering Problems

A new method based on machine learning for the FDTD method

Two-Dimensional Electromagnetic Solver Based on Deep Learning Technique

Machine Learning Based FDTD Method for 3D Electromagnetic Simulation

Modeling of Ground Penetrating Radar System Based on Recurrent Neural Network Architecture

Recurrent Neural Network-Assisted Truncation of Convolutional Perfectly Matched Layers for FDTD

Modeling EM Problem with Deep Neural Networks

Neural network EM-Field based modeling for 3d substructure in finite element method

Differentiable-Decision-Tree-Based Neural Turing Machine Model Integrated Into FDTD for Implementing EM Problems

Application of Deep Learning in FDTD Method

Reactive Near-Field to 3-Meter Far-Field Transformation Based on Deep Convolutional Neural Networks and Plane Wave Spectrum

A Time Domain Neural Network Method for Scattering of Complex Scatterers

The Equivalence and Realization of Neural Network and Finite Differences Time Domain

A Maxwell's Equations Based Deep Learning Method for Time Domain Electromagnetic Simulations

An Efficient FDTD-CNN Method for Analyzing the Time-Domain Response of Objects above Layered Half-Space under HPEMP Illumination