Low Photon Count Phase Retrieval Using Deep Learning

Abstract

Imaging systems' performance at low light intensity is affected by shot noise, which becomes increasingly strong as the power of the light source decreases. In this paper we experimentally demonstrate the use of deep neural networks to recover objects illuminated with weak light and demonstrate better performance than with the classical Gerchberg-Saxton phase retrieval algorithm for equivalent signal over noise ratio. Prior knowledge about the object is implicitly contained in the training data set and feature detection is possible for a signal over noise ratio close to one. We apply this principle to a phase retrieval problem and show successful recovery of the object's most salient features with as little as one photon per detector pixel on average in the illumination beam. We also show that the phase reconstruction is significantly improved by training the neural network with an initial estimate of the object, as opposed as training it with the raw intensity measurement.

8 pages, 5 figures

References in corpus (6)

• Phase recovery and holographic image reconstruction using deep learning in neural networks
• Efficient and accurate inversion of multiple scattering with deep learning
• Convolutional neural networks that teach microscopes how to image
• Phase retrieval for Fourier Ptychography under varying amount of measurements
• Imaging through glass diffusers using densely connected convolutional networks
• Propagation based phase retrieval of simulated intensity measurements using artificial neural networks

Cited by in corpus (46)

• On the use of deep learning for phase recovery
• Machine Learning on Neutron and X-Ray Scattering
• Fourier Imager Network (FIN): A deep neural network for hologram reconstruction with superior external generalization
• Self-supervised learning of hologram reconstruction using physics consistency
• Intelligent Meta-Imagers: From Compressed to Learned Sensing
• Holographic image reconstruction with phase recovery and autofocusing using recurrent neural networks
• High-Resolution Limited-Angle Phase Tomography of Dense Layered Objects Using Deep Neural Networks
• Multi-scale Convolutional Neural Networks for Inverse Problems
• Molecular rotation movie filmed with high-harmonic generation
• Accurate phase retrieval of complex point spread functions with deep residual neural networks
• A Two-step-training Deep Learning Framework for Real-time Computational Imaging without Physics Priors
• Adaptive 3D convolutional neural network-based reconstruction method for 3D coherent diffraction imaging
• On the interplay between physical and content priors in deep learning for computational imaging
• Ghost imaging based on Y-net: a dynamic coding and conjugate-decoding approach
• Spatial images from temporal data
• Blending Diverse Physical Priors with Neural Networks
• eFIN: Enhanced Fourier Imager Network for generalizable autofocusing and pixel super-resolution in holographic imaging
• Optimal Physical Preprocessing for Example-Based Super-Resolution
• Few-shot Transfer Learning for Holographic Image Reconstruction using a Recurrent Neural Network
• Limited-angle tomographic reconstruction of dense layered objects by dynamical machine learning
• Power-law Scaling to Assist with Key Challenges in Artificial Intelligence
• Robust and Efficient Single-Pixel Image Classificationwith Nonlinear Optics
• Deep Learning-Assisted Classification of Site-Resolved Quantum Gas Microscope Images
• Calibration-free quantitative phase imaging using data-driven aberration modeling
• Single-Pixel Pattern Recognition with Coherent Nonlinear Optics
• Low-intensity illumination for lensless digital holographic microscopy with minimized sample interaction
• Physics-assisted Generative Adversarial Network for X-Ray Tomography
• Learning to synthesize: splitting and recombining low and high spatial frequencies for image recovery
• Signal retrieval with measurement system knowledge using variational generative model
• Attentional Ptycho-Tomography (APT) for three-dimensional nanoscale X-ray imaging with minimal data acquisition and computation time
• PDMM: A novel Primal-Dual Majorization-Minimization algorithm for Poisson Phase-Retrieval problem
• Fourier Phase Retrieval with Extended Support Estimation via Deep Neural Network
• Deep-learning real-time phase retrieval of imperfect diffraction patterns from X-ray free-electron lasers
• Noise-robust latent vector reconstruction in ptychography using deep generative models
• Phase Retrieval with Holography and Untrained Priors: Tackling the Challenges of Low-Photon Nanoscale Imaging
• Investigating the robustness of a learning-based method for quantitative phase retrieval from propagation-based x-ray phase contrast measurements under laboratory conditions
• Towards smart optical focusing: Deep learning-empowered wavefront shaping in nonstationary scattering media
• Deep Shape from Polarization
• Phase retrieval via Zernike phase contrast microscopy with an untrained neural network
• Snapshot Interferometric 3D Imaging by Compressive Sensing and Deep Learning
• Deep learning based high-resolution incoherent x-ray imaging with a single-pixel detector
• 3D-2D Neural Nets for Phase Retrieval in Noisy Interferometric Imaging
• Computer-free, all-optical reconstruction of holograms using diffractive networks
• Neural networks with quantum states of light
• Phase retrieval from 4-dimensional electron diffraction datasets
• Blaschke Product Neural Networks (BPNN): A Physics-Infused Neural Network for Phase Retrieval of Meromorphic Functions