Coordinate-based Neural Network for Fourier Phase Retrieval

TL;DR

This paper introduces SCAN, a coordinate neural network that improves Fourier phase retrieval by integrating amplitude and phase predictions, outperforming traditional methods especially under noisy conditions and in ptychography applications.

Contribution

The paper presents a novel unsupervised coordinate neural network, SCAN, that enhances phase retrieval accuracy and noise robustness by jointly modeling amplitude and phase.

Findings

SCAN outperforms traditional iterative methods in accuracy.

SCAN demonstrates superior noise robustness.

SCAN excels in ptychography applications.

Abstract

Fourier phase retrieval is essential for high-definition imaging of nanoscale structures across diverse fields, notably coherent diffraction imaging. This study presents the Single impliCit neurAl Network (SCAN), a tool built upon coordinate neural networks meticulously designed for enhanced phase retrieval performance. Remedying the drawbacks of conventional iterative methods which are easiliy trapped into local minimum solutions and sensitive to noise, SCAN adeptly connects object coordinates to their amplitude and phase within a unified network in an unsupervised manner. While many existing methods primarily use Fourier magnitude in their loss function, our approach incorporates both the predicted magnitude and phase, enhancing retrieval accuracy. Comprehensive tests validate SCAN's superiority over traditional and other deep learning models regarding accuracy and noise robustness. We also demonstrate that SCAN excels in the ptychography setting.