Self-supervised physics-informed generative networks for phase retrieval from a single X-ray hologram

2025-08-21

AI Summary

In X-ray phase-contrast imaging, quantitative phase retrieval from a single hologram is hindered by the inherent loss of phase information in intensity measurements. To address this, we propose a self-supervised, physics-informed generative adversarial network (PIGAN) that jointly reconstructs phase and absorption distributions directly from a single intensity image—without requiring paired ground-truth or synthetic training data. Our method embeds the Fresnel diffraction model into the loss function, synergistically integrating physical constraints, adversarial learning, and self-supervision to eliminate reliance on empirical priors or manual hyperparameter tuning. Extensive validation on both simulated data and experimental synchrotron data from PETRA III demonstrates strong generalizability across diverse samples and imaging conditions. The reconstructions are quantitatively accurate, robust, and stable. To our knowledge, this is the first end-to-end unsupervised framework for quantitative phase retrieval from a single X-ray hologram under explicit physical constraints.

Technology Category

• Computer Vision: Generative Adversarial Networks (GANs) for Vision
• Machine Learning: Unsupervised & Self-Supervised Learning
• Humans and AI: Explainable AI (XAI) for Human Understanding

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Security and Privacy: Data transparency and provenance

Abstract

X-ray phase contrast imaging significantly improves the visualization of structures with weak or uniform absorption, broadening its applications across a wide range of scientific disciplines. Propagation-based phase contrast is particularly suitable for time- or dose-critical in vivo/in situ/operando (tomography) experiments because it requires only a single intensity measurement. However, the phase information of the wave field is lost during the measurement and must be recovered. Conventional algebraic and iterative methods often rely on specific approximations or boundary conditions that may not be met by many samples or experimental setups. In addition, they require manual tuning of reconstruction parameters by experts, making them less adaptable for complex or variable conditions. Here we present a self-learning approach for solving the inverse problem of phase retrieval in the near-field regime of Fresnel theory using a single intensity measurement (hologram). A physics-informed generative adversarial network is employed to reconstruct both the phase and absorbance of the unpropagated wave field in the sample plane from a single hologram. Unlike most deep learning approaches for phase retrieval, our approach does not require paired, unpaired, or simulated training data. This significantly broadens the applicability of our approach, as acquiring or generating suitable training data remains a major challenge due to the wide variability in sample types and experimental configurations. The algorithm demonstrates robust and consistent performance across diverse imaging conditions and sample types, delivering quantitative, high-quality reconstructions for both simulated data and experimental datasets acquired at beamline P05 at PETRA III (DESY, Hamburg), operated by Helmholtz-Zentrum Hereon. Furthermore, it enables the simultaneous retrieval of both phase and absorption information.

Problem

• Recovering lost phase information from single X-ray hologram measurements
• Overcoming limitations of conventional methods requiring specific approximations
• Eliminating need for paired training data in phase retrieval

Innovation

• Self-supervised physics-informed generative adversarial network
• Single hologram phase retrieval without training data
• Simultaneous reconstruction of phase and absorption