Image Sensors Based on Nonlinear Optical Neural Networks

We use a multilayer, nonlinear optical neural network as a pre-processor for machine vision. On tasks including identification of real objects illuminated by broadband incoherent light, we show that nonlinear optical pre-processing outperforms linear pre-processing.

Saved in:
Bibliographic Details
Published in2023 Conference on Lasers and Electro-Optics (CLEO) pp. 1 - 2
Main Authors Sohoni, Mandar M., Wang, Tianyu, Wright, Logan G., Stein, Martin M., Ma, Shi-Yuan, Onodera, Tatsuhiro, Anderson, Maxwell G., McMahon, Peter L.
Format Conference Proceeding
LanguageEnglish
Published OSA 01.05.2023
Subjects
Lasers and electrooptics
Neural networks
Nonhomogeneous media
Optical computing
Optical fiber networks
Optical imaging
Optical sensors
Online AccessGet full text

Cover

More Information
Summary:We use a multilayer, nonlinear optical neural network as a pre-processor for machine vision. On tasks including identification of real objects illuminated by broadband incoherent light, we show that nonlinear optical pre-processing outperforms linear pre-processing.
DOI:10.1364/CLEO_AT.2023.STu4F.3