Terahertz Metamaterial Intelligent Identification by Private Preserving Deep Learning

Metamaterials are artificial materials, designed with special structures, and can exhibit controllable electromagnetic properties. To address the looming threat of stealthy monitoring caused by metamaterials, the detection and identification of them is the next research frontier but have not yet been explored. In this research, a crypto-oriented convolutional neural network is designed to identify metamaterial in mixtures, which is an accurate and secure paradigm of AI-based terahertz smart sensing.