A Frequency-Multiplexed Function-Approximation Coherent Neural Networks To Learn Phase Values by Use of Volume Hologram
Amornrat Limmanee, Sotaro Kawata, Akira Hirose · Frontiers in Optics · 2005
We propose a frequency-multiplexed function-approximation coherent neural network to learn optical phase values by use of volume hologram. Experiments demonstrate that desired phase values are obtained as the output of the coherent neural network system.