Optical convolutional neural network based on an amplitude modulation Spatial Light Modulator and a 4-level phase plate

A method to realize complex amplitude modulation with a 4-level phase plate and an amplitude-modulated SLM is proposed in this paper. We implement this complex amplitude modulation system in classical 4f system and use this system to realize highly precise convolution to verify its feasibility. We also incorporate this system with a one-layer Convolutional Neural Network (CNN) as its convolutional layer to accomplish a handwriting classification inference. The image classification accuracy is comparable to all-electrical result with the same framework.

Automated summary

This paper proposes a method to achieve complex amplitude modulation using a 4-level phase plate and an amplitude-modulated SLM. The system is implemented in a classical 4f system and used for highly precise convolution, as well as integrated with a one-layer Convolutional Neural Network for handwriting classification inference. The image classification accuracy is comparable to the all-electrical result within the same framework.