Diffractive Deep Neural Network for Microwave Imaging Based on a Multi-Layer Transmitarray Antenna
Daoqi Liu, Tao Shan, Maokun Li, Fan Yang, Shenheng Xu · 2023
In this study, we designed a diffractive deep neural network (D2NN) structure using a multi-layer transmitarray antenna. Each transmitarray surface is treated as a physical neural network (NN) layer, with each element representing a neuron. The interaction between electromagnetic (EM) waves and transmitarray surfaces can be regarded as the computation of NNs. The proposed D2NN is applied for microwave imaging of the metal targets. Numerical results demonstrate that D2NN can provide reliable imaging with good similarity to the target in terms of size, shape, and position.