Inpainting Galactic Foreground Intensity and Polarization maps using Convolutional Neural Network

Abstract

Deep convolutional neural networks have been a popular tool for image generation and restoration. The performance of these networks is related to the capability of learning realistic features from a large dataset. In this work, we applied the problem of inpainting non-Gaussian signal, in the context of Galactic diffuse emissions at the millimetric and sub-millimetric regimes, specifically Synchrotron and Thermal Dust emission. Both of them are affected by contamination at small angular scales due to extra-galactic radio sources (the former) and to dusty star-forming galaxies (the latter). We consider the performances of a nearest-neighbors inpainting technique and compare it with two novels methodologies relying on generative Neural Networks. We show that the generative network is able to reproduce the statistical properties of the ground truth signal more consistently with high confidence level. The Python Inpainter for Cosmological and AStrophysical SOurces (PICASSO) is a package encoding a suite of inpainting methods described in this work and has been made publicly available.

References in corpus (34)

• HEALPix -- a Framework for High Resolution Discretization, and Fast Analysis of Data Distributed on the Sphere
• The Simons Observatory: Science goals and forecasts
• Signature of Gravity Waves in Polarization of the Microwave Background
• Deep Image Prior
• Planck 2015 results. XVII. Constraints on primordial non-Gaussianity
• BICEP2 / Keck Array x: Constraints on Primordial Gravitational Waves using Planck, WMAP, and New BICEP2/Keck Observations through the 2015 Season
• Planck 2013 Results. XXIV. Constraints on primordial non-Gaussianity
• CMB Anisotropies: Total Angular Momentum Method
• Planck 2018 results. IV. Diffuse component separation
• A unified pseudo-$C_{\ell }$ framework
• The Python Sky Model: software for simulating the Galactic microwave sky
• The Atacama Cosmology Telescope: Two-Season ACTPol Spectra and Parameters
• PiiGAN: Generative Adversarial Networks for Pluralistic Image Inpainting
• A Measurement of the Cosmic Microwave Background $B$-Mode Polarization Power Spectrum at Sub-Degree Scales from 2 years of POLARBEAR Data
• CosmoGAN: creating high-fidelity weak lensing convergence maps using Generative Adversarial Networks
• The State-of-Play of Anomalous Microwave Emission (AME) Research
• Measurements of B-mode Polarization of the Cosmic Microwave Background from 500 Square Degrees of SPTpol Data
• The S-PASS view of polarized Galactic Synchrotron at 2.3 GHz as a contaminant to CMB observations
• Fast cosmic web simulations with generative adversarial networks
• The Atacama Cosmology Telescope: Extragalactic Sources at 148 GHz in the 2008 Survey
• S-band Polarization All Sky Survey (S-PASS): survey description and maps
• DeepCMB: Lensing Reconstruction of the Cosmic Microwave Background with Deep Neural Networks
• The Atacama Cosmology Telescope: Two-season ACTPol Extragalactic Point Sources and their Polarization properties
• Convolutional Neural Networks on the HEALPix sphere: a pixel-based algorithm and its application to CMB data analysis
• Filling in CMB map missing data using constrained Gaussian realizations
• Impact of polarised extragalactic sources on the measurement of CMB B-mode anisotropies
• Magnetic field disorder and Faraday effects on the polarization of extragalactic radio sources
• Low-l CMB Analysis and Inpainting
• Forecasting Polarized Radio Sources for CMB observations
• Fractional Polarisation of Extragalactic Sources in the 500-square-degree SPTpol Survey
• Cleaning our own Dust: Simulating and Separating Galactic Dust Foregrounds with Neural Networks
• A 3D model for CO molecular line emission as a potential CMB polarization contaminant
• Foreground model recognition through Neural Networks for CMB B-mode observations
• How to avoid X'es around point sources in maximum likelihood CMB maps

Cited by in corpus (16)

• Lossless, Scalable Implicit Likelihood Inference for Cosmological Fields
• Machine Learning for Observational Cosmology
• ForSE: a GAN based algorithm for extending CMB foreground models to sub-degree angular scales
• Recovering the CMB Signal with Machine Learning
• MillimeterDL: Deep Learning Simulations of the Microwave Sky
• Simulated catalogs and maps of radio galaxies at millimeter wavelengths in Websky
• Encoding large scale cosmological structure with Generative Adversarial Networks
• Inpainting CMB maps using Partial Convolutional Neural Networks
• The Andromeda Gamma-Ray Excess: Background Systematics of the Millisecond Pulsars and Dark Matter Interpretations
• Inpainting hydrodynamical maps with deep learning
• CMB-PAInT: An inpainting tool for the cosmic microwave background
• Self-supervised component separation for the extragalactic submillimeter sky
• A convolutional-neural-network estimator of CMB constraints on dark matter energy injection
• Constraining the polarisation flux density and angle of point sources by training a convolutional neural network
• Extending Galactic foreground emission with neural networks
• Removing the mask -- reconstructing a scalar field on the sphere from a masked field