High-fidelity demodulation of vortex beams through dynamic scattering media using a physically constrained deep...

Area of Science:

• Optics and Photonics
• Machine Learning
• Biomedical Imaging

Background:

• Vortex beams with orbital angular momentum (OAM) are crucial for high-capacity optical communication and imaging.
• Wavefront disruption by multiple scattering in dynamic media, like biological tissues, limits OAM beam applications.
• Brownian motion in such media causes field decorrelation, rendering conventional decoding methods ineffective.

Purpose of the Study:

• To develop a robust and high-fidelity deep-learning demodulation technique for OAM vortex beams scattered in dynamic media.
• To integrate data-driven approaches with physical principles for improved decoding accuracy and interpretability.
• To address the challenge of wavefront distortion caused by Brownian motion in scattering environments.

Main Methods:

• Utilized experimentally acquired full-field speckle patterns from a dynamic aqueous milk scattering system for training.
• Introduced a quantum-limited-fidelity residual network (QLF-ResNet) incorporating a forced L2 normalization layer for energy conservation.
• Applied rotation-based data augmentation for comprehensive end-to-end model training.

Main Results:

• Achieved an average classification accuracy of 91.25% ± 4.15% for OAM modes l = 1-4.
• Successfully resolved complex coefficients, suppressed crosstalk between OAM modes, and accurately rendered beam intensity and phase.
• Demonstrated superior performance compared to pure data-driven models by avoiding common artifacts.

Conclusions:

• The proposed deep-learning method offers a robust and interpretable decoding scheme for OAM vortex beams in complex, time-varying scattering environments.
• Integrating physical priors, such as energy conservation, enhances model reliability and performance.
• This technique paves the way for advanced applications in optical communication and imaging through dynamic biological tissues.