Machine-learning reprogrammable metasurface imager

Lianlin Li,Tie Jun Cui,Ya Shuang,Cheng-Wei Qiu,Che Liu,Hengxin Ruan,Ying Li,Andrea Alù

doi:10.1038/s41467-019-09103-2

Abstract

Conventional microwave imagers usually require either time-consuming data acquisition, or complicated reconstruction algorithms for data post-processing, making them largely ineffective for complex in-situ sensing and monitoring. Here, we experimentally report a real-time digital-metasurface imager that can be trained in-situ to generate the radiation patterns required by machine-learning optimized measurement modes. This imager is electronically reprogrammed in real time to access the optimized solution for an entire data set, realizing storage and transfer of full-resolution raw data in dynamically varying scenes. High-accuracy image coding and recognition are demonstrated in situ for various image sets, including hand-written digits and through-wall body gestures, using a single physical hardware imager, reprogrammed in real time. Our electronically controlled metasurface imager opens new venues for intelligent surveillance, fast data acquisition and processing, imaging at various frequencies, and beyond.

Highlights

Conventional microwave imagers usually require either time-consuming data acquisition, or complicated reconstruction algorithms for data post-processing, making them largely ineffective for complex in-situ sensing and monitoring
Each row of H corresponds to one measurement mode, and the number of rows equals the number of measurements
In order to manipulate arbitrarily the measurement modes needed by machine learning with high accuracy in real time, we propose an electrically modulated 2-bit reprogrammable coding metasurface

Summary

Introduction

Conventional microwave imagers usually require either time-consuming data acquisition, or complicated reconstruction algorithms for data post-processing, making them largely ineffective for complex in-situ sensing and monitoring. In many cases, such as for security screening or pipeline monitoring, the flux of data is so large that full-resolution imaging is inefficient and unmanageable, and represents a huge waste of resources and energy, since only a few properties of the images are of interest, such as the position of an object, or its dynamic changes[8,9,10,11,12,13] These situations require an imaging device that can instantly reconstruct scenes in an intelligent and efficient way, i.e., rendering the important feature extraction with high speed, fidelity, and compression ratio, as it commonly happens in biological systems, such as our brain. Reprogrammable digital coding metamaterials have been proposed to dynamic holography[18], compressive imaging[19], beam scan[20], and other related operations[20], in which the coding sequence of the meta-atoms with digitized states have been used to manipulate the impinging microwave waves in an efficient and controllable fashion

Methods

Results

Conclusion