Analysis of Diffractive Neural Networks for Seeing Through Random Diffusers

TL;DR

This paper analyzes all-optical diffractive neural networks for imaging through unknown diffusers, exploring training strategies, generalization, and robustness to improve imaging in complex media without digital computation.

Contribution

It provides a comprehensive analysis of training methods, network design, and robustness strategies for diffractive neural networks to see through random diffusers.

Findings

Using multiple diffusers during training improves generalization.

Additional diffractive layers enhance robustness to new diffusers.

Deliberate misalignments during training increase network resilience.

Abstract

Imaging through diffusive media is a challenging problem, where the existing solutions heavily rely on digital computers to reconstruct distorted images. We provide a detailed analysis of a computer-free, all-optical imaging method for seeing through random, unknown phase diffusers using diffractive neural networks, covering different deep learning-based training strategies. By analyzing various diffractive networks designed to image through random diffusers with different correlation lengths, a trade-off between the image reconstruction fidelity and distortion reduction capability of the diffractive network was observed. During its training, random diffusers with a range of correlation lengths were used to improve the diffractive network's generalization performance. Increasing the number of random diffusers used in each epoch reduced the overfitting of the diffractive network's imaging performance to known diffusers. We also demonstrated that the use of additional diffractive layers improved the generalization capability to see through new, random diffusers. Finally, we introduced deliberate misalignments in training to 'vaccinate' the network against random layer-to-layer shifts that might arise due to the imperfect assembly of the diffractive networks. These analyses provide a comprehensive guide in designing diffractive networks to see through random diffusers, which might profoundly impact many fields, such as biomedical imaging, atmospheric physics, and autonomous driving.