Lightweight Metasurface Absorber Customization with a Conditional Generative Adversarial Network

Deep learning algorithm known as conditional generative adversarial network (cGAN), which enables the generation of novel graphical patterns. Here we propose a deep learning method based on cGAN for rapid customization of lightweight metasurface absorbers. As a result, the proposed method offers design flexibility and diversity, facilitating the rapid creation of compliant metasurface designs. Moreover, this approach can be readily applied to the design of other on-demand designs.