Learning neural representations for X-ray ptychography reconstruction with unknown probes

2025-09-04

AI Summary

X-ray ptychography holds great promise for nanoscale imaging, yet conventional iterative and deep learning methods suffer substantial degradation in reconstruction accuracy under unknown probe conditions and low signal-to-noise ratios (e.g., low-dose or high-speed acquisition). To address this, we propose PtyINR—a self-supervised framework that jointly models both the object and the probe as continuous implicit neural representations (INRs), the first of its kind for ptychography. Without requiring prior knowledge of the probe, PtyINR enables end-to-end joint optimization via physics-constrained self-supervised learning, directly recovering both the object structure and probe distribution from raw diffraction patterns. Extensive evaluations on both synthetic and experimental datasets demonstrate that PtyINR consistently outperforms state-of-the-art methods, particularly excelling in low-signal regimes by significantly improving reconstruction fidelity and robustness. This work establishes a new paradigm for high-throughput, low-dose X-ray nanoscale imaging.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Machine Learning: Unsupervised & Self-Supervised Learning
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Security and Privacy: Data transparency and provenance

Abstract

X-ray ptychography provides exceptional nanoscale resolution and is widely applied in materials science, biology, and nanotechnology. However, its full potential is constrained by the critical challenge of accurately reconstructing images when the illuminating probe is unknown. Conventional iterative methods and deep learning approaches are often suboptimal, particularly under the low-signal conditions inherent to low-dose and high-speed experiments. These limitations compromise reconstruction fidelity and restrict the broader adoption of the technique. In this work, we introduce the Ptychographic Implicit Neural Representation (PtyINR), a self-supervised framework that simultaneously addresses the object and probe recovery problem. By parameterizing both as continuous neural representations, PtyINR performs end-to-end reconstruction directly from raw diffraction patterns without requiring any pre-characterization of the probe. Extensive evaluations demonstrate that PtyINR achieves superior reconstruction quality on both simulated and experimental data, with remarkable robustness under challenging low-signal conditions. Furthermore, PtyINR offers a generalizable, physics-informed framework for addressing probe-dependent inverse problems, making it applicable to a wide range of computational microscopy problems.

Problem

• Reconstructing images with unknown probes in X-ray ptychography
• Overcoming limitations of conventional methods under low-signal conditions
• Addressing probe-dependent inverse problems in computational microscopy

Innovation

• Self-supervised neural representation framework for ptychography
• End-to-end reconstruction from raw diffraction patterns
• Continuous parameterization of both object and probe