Deep learning sheds new light on non-orthogonal optical multiplexing.

Published in: Light, science & applications (2024)

A deep neural network for non-orthogonal input channel encoding is proposed to recover speckle images through a multimode fiber. This novel approach could shed new light on the non-orthogonal optical multiplexing over a scattering medium.

Keyphrases

deep learning
neural network
high resolution
convolutional neural network
high speed
artificial intelligence
machine learning
mass spectrometry