Low Photon Budget Phase Retrieval with Perceptual Loss Trained Deep Neural Networks

TL;DR

This paper demonstrates that using a perceptual loss function in deep neural networks enhances phase retrieval quality in low photon count scenarios, outperforming previous methods that used correlation-based loss functions.

Contribution

It introduces the use of perceptual loss functions for DNN-based phase retrieval under photon-starved conditions, improving reconstruction quality.

Findings

Perceptual loss improves phase retrieval reconstructions.

Deep neural networks outperform classical methods in low photon regimes.

Enhanced robustness to Poisson noise in phase retrieval tasks.

Abstract

Deep neural networks (DNNs) are efficient solvers for ill-posed problems and have been shown to outperform classical optimization techniques in several computational imaging problems. DNNs are trained by solving an optimization problem implies the choice of an appropriate loss function, i.e., the function to optimize. In a recent paper [A. Goy \textit{et al.}, Phys. Rev. Lett. 121(24), 243902 (2018)], we showed that DNNs trained with the negative Pearson correlation coefficient as the loss function are particularly fit for photon-starved phase retrieval problems, which are fundamentally affected by strong Poison noise. In this paper we demonstrate that the use of a perceptual loss function significantly improves the reconstructions.