[16p-M_110-1~13] 3.2 Information photonics and image engineering

5:30 PM - 5:45 PM JST(8:30 AM - 8:45 AM UTC)Presentation by Applicant for JSAP Young Scientists Presentation Award

[16p-M_110-13] Design of 2D Metasurface Lenses for Specialized Lighting with Machine Learning

〇(B)Toma Unno¹, Gai Ichisawa¹, Towa Maekawa¹, Tomohiro Amemiya¹ (1.Science Tokyo)

Keywords:

metasurface lenses,metalens,machine learning

We propose an efficient design method using machine learning to reduce the design load of metalenses for special illumination requiring complex light distribution control. We constructed a neural network that inversely estimates the required angular modulation characteristics from a target illuminance distribution, using training data based on a geometric optics model. Based on these estimates, nanostructure parameters were determined using Rigorous Coupled-Wave Analysis (RCWA). We report design results that reproduce the target light distributions across visible wavelengths.