Kyujung Kim

Jul 10, 2024 | opg.optica.org | Kyujung Kim |Mechatronics Engineering

The conventional design process for metasurfaces is time-consuming andcomputationally expensive. To address this challenge, we utilize adeep convolutional generative adversarial network (DCGAN) to generatenew nanohole metastructure designs that match a desired transmittancespectrum in the visible range. The trained DCGAN model demonstrates anexceptional performance in generating diverse and manufacturablemetastructure designs that closely resemble the target opticalproperties.