Phase retrieval via Zernike phase contrast microscopy with an untrained neural network

TL;DR

This paper introduces an untrained neural network approach as a prior for phase retrieval in Zernike phase contrast microscopy, eliminating manual regularization and improving accuracy and robustness in phase reconstruction.

Contribution

It extends existing phase retrieval methods by using an untrained neural network as a prior, removing the need for empirical regularization tuning.

Findings

Improved accuracy in phase retrieval over existing methods

Enhanced robustness demonstrated through numerical and experimental data

Feasibility of deep priors for phase retrieval in incoherent illumination

Abstract

Zernike's phase contrast microscopy (PCM) is among the most widely used techniques for observing phase objects, but it lacks quantitative nature, as it cannot directly provide phase information. Current methods for computationally extracting phase distributions from PCM images, however, rely heavily on empirical regularization parameter tuning. In this paper we extend an existing approach by employing an untrained neural network as an image prior, removing the need for manual regularization. We quantitatively demonstrate improved accuracy and robustness in phase retrieval compared to existing methods, using numerical and experimental PCM images. Our results confirm the feasibility of applying deep priors for phase retrieval in incoherent illumination setups.