Reconstruction of fractional vortex phase evolution by generative adversarial networks

Area of Science:

• Optical communication
• Structured light
• Digital signal coding

Background:

• Orbital angular momentum (OAM) and vortex optical phase are crucial for optical communication.
• Accurate vortex optical phase recovery is key to enhancing communication coding efficiency.

Purpose of the Study:

• To propose a deep learning model for accurate phase recovery of fractional vortex patterns.
• To address the challenge of phase evolution affecting fractional vortex patterns at different transmission distances.

Main Methods:

• A deep learning model based on a generative adversarial network (GAN) was developed.
• The model accurately recovers phase image information of fractional vortex patterns at any diffraction distance.

Main Results:

• The GAN-based phase recovery is independent of diffraction distance.
• This represents the first application of GANs to fractional-order optical vortices.
• The method improves upon other deep learning approaches for phase recovery.

Conclusions:

• The proposed GAN model offers a novel approach for accurate identification of multi-singular structured light.
• This work enhances the efficiency and accuracy of optical communication coding.