Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Data driven drift correction for complex optical systems.

Journal of synchrotron radiation·2026
Same author

High-Throughput Exploration of Lithium-Alloy Protection Layers for High-Performance Lithium-Metal Batteries.

ACS applied energy materials·2025
Same author

Ionic tuning of cobaltites at the nanoscale.

Physical review materials·2024
Same author

Active Learning for Rapid Targeted Synthesis of Compositionally Complex Alloys.

Materials (Basel, Switzerland)·2024
Same author

MLExchange: A web-based platform enabling exchangeable machine learning workflows for scientific studies.

Annual Workshop on Extreme-scale Experiment-in-the-Loop Computing : XLOOP. Annual Workshop on Extreme-scale Experiment-in-the-Loop Computing·2023
Same author

Tuning In-Plane Magnetic Anisotropy and Interfacial Exchange Coupling in Epitaxial La<sub>2/3</sub>Sr<sub>1/3</sub>CoO<sub>3</sub>/La<sub>2/3</sub>Sr<sub>1/3</sub>MnO<sub>3</sub> Heterostructures.

ACS applied materials & interfaces·2023

Related Experiment Video

Updated: Jul 7, 2025

Digital Inline Holographic Microscopy DIHM of Weakly-scattering Subjects
10:16

Digital Inline Holographic Microscopy DIHM of Weakly-scattering Subjects

Published on: February 8, 2014

12.3K

Physics constrained unsupervised deep learning for rapid, high resolution scanning coherent diffraction

Oliver Hoidn1, Aashwin Ananda Mishra2, Apurva Mehta2

  • 1SLAC National Accelerator Laboratory, Menlo Park, CA, USA. ohoidn@slac.stanford.edu.

Scientific Reports
|December 20, 2023
PubMed
Summary
This summary is machine-generated.

We developed PtychoPINN, an unsupervised deep learning method for faster and higher-quality imaging. This physics-informed neural network improves resolution and reconstruction speed for applications like X-ray imaging.

More Related Videos

Compact Lens-less Digital Holographic Microscope for MEMS Inspection and Characterization
10:28

Compact Lens-less Digital Holographic Microscope for MEMS Inspection and Characterization

Published on: July 5, 2016

10.3K
High-resolution, High-speed, Three-dimensional Video Imaging with Digital Fringe Projection Techniques
11:34

High-resolution, High-speed, Three-dimensional Video Imaging with Digital Fringe Projection Techniques

Published on: December 3, 2013

15.7K

Related Experiment Videos

Last Updated: Jul 7, 2025

Digital Inline Holographic Microscopy DIHM of Weakly-scattering Subjects
10:16

Digital Inline Holographic Microscopy DIHM of Weakly-scattering Subjects

Published on: February 8, 2014

12.3K
Compact Lens-less Digital Holographic Microscope for MEMS Inspection and Characterization
10:28

Compact Lens-less Digital Holographic Microscope for MEMS Inspection and Characterization

Published on: July 5, 2016

10.3K
High-resolution, High-speed, Three-dimensional Video Imaging with Digital Fringe Projection Techniques
11:34

High-resolution, High-speed, Three-dimensional Video Imaging with Digital Fringe Projection Techniques

Published on: December 3, 2013

15.7K

Area of Science:

  • Optics and Imaging Science
  • Computational Physics
  • Machine Learning Applications

Background:

  • Coherent diffractive imaging (CDI) and ptychography overcome optical resolution limits.
  • Iterative phase recovery in CDI and ptychography is time-consuming, hindering real-time applications.
  • Supervised deep learning accelerates reconstruction but compromises image quality and requires extensive labeled data.

Purpose of the Study:

  • To introduce PtychoPINN, an unsupervised physics-informed neural network for rapid and high-fidelity image reconstruction.
  • To address the limitations of existing methods, including speed, image quality, and data requirements.

Main Methods:

  • Developed PtychoPINN, an unsupervised physics-informed neural network.
  • Combined diffraction forward map with real-space constraints from overlapping measurements.
  • Leveraged unsupervised learning to avoid the need for labeled training data.

Main Results:

  • Achieved a 100-to-1000 speedup in reconstruction compared to traditional methods.
  • Improved linear resolution by a factor of 4.
  • Enhanced image quality with an 8 dB improvement in Peak Signal-to-Noise Ratio (PSNR).
  • Demonstrated improved generalizability and robustness of the reconstruction method.

Conclusions:

  • PtychoPINN offers a blend of computational efficiency and high performance for real-time imaging.
  • The method shows significant promise for high-throughput applications at facilities like X-ray free electron lasers (XFELs).
  • PtychoPINN advances the field of high-resolution imaging by overcoming current speed and quality trade-offs.