Phase Retrieval using Single-Instance Deep Generative Prior

TL;DR

This paper introduces a new phase retrieval method leveraging a single-instance deep generative prior, effectively handling complex-valued crystal data where existing methods struggle without precise support information.

Contribution

The paper presents a novel phase retrieval approach using a single-instance deep generative prior, improving performance on complex-valued data without requiring precise support info.

Findings

Effective on complex-valued crystal data

Outperforms existing methods without precise support

Works well on realistic data

Abstract

Several deep learning methods for phase retrieval exist, but most of them fail on realistic data without precise support information. We propose a novel method based on single-instance deep generative prior that works well on complex-valued crystal data.