期刊:Conference on Lasers and Electro-Optics日期:2019-01-01被引量:1
标识
DOI:10.1364/cleo_si.2019.sf1n.2
摘要
We report deep learning-based design of diffractive neural networks. Following its fabrication, a diffractive neural network can all-optically perform user-defined tasks through diffraction. We demonstrate the applications of this framework for classification and imaging tasks.