Generative adversarial networks recover features in astrophysical images of galaxies beyond the deconvolution limit
Reads0
Chats0
TLDR
The ability to better recover detailed features from low-signal-to-noise and low angular resolution imaging data significantly increases the ability to study existing data sets of astrophysical objects as well as future observations with observatories such as the Large Synoptic Sky Telescope and the Hubble and James Webb space telescopes.Abstract:
Observations of astrophysical objects such as galaxies are limited by various sources of random and systematic noise from the sky background, the optical system of the telescope and the detector used to record the data. Conventional deconvolution techniques are limited in their ability to recover features in imaging data by the Shannon-Nyquist sampling theorem. Here we train a generative adversarial network (GAN) on a sample of $4,550$ images of nearby galaxies at $0.01<z<0.02$ from the Sloan Digital Sky Survey and conduct $10\times$ cross validation to evaluate the results. We present a method using a GAN trained on galaxy images that can recover features from artificially degraded images with worse seeing and higher noise than the original with a performance which far exceeds simple deconvolution. The ability to better recover detailed features such as galaxy morphology from low-signal-to-noise and low angular resolution imaging data significantly increases our ability to study existing data sets of astrophysical objects as well as future observations with observatories such as the Large Synoptic Sky Telescope (LSST) and the Hubble and James Webb space telescopes.read more
Citations
More filters
Journal ArticleDOI
Probabilistic Random Forest: A Machine Learning Algorithm for Noisy Data Sets
TL;DR: In this paper, the authors modify the long-established Random Forest (RF) algorithm to take into account uncertainties in measurements (i.e., features) as well as in assigned classes (e.g., labels).
Journal ArticleDOI
Image-Based Model Parameter Optimization Using Model-Assisted Generative Adversarial Networks
TL;DR: This work proposes and demonstrates the use of a model-assisted generative adversarial network to produce fake images that accurately match true images through the variation of the parameters of the model that describes the features of the images.
Journal ArticleDOI
Deep generative models for galaxy image simulations
François Lanusse,Rachel Mandelbaum,Siamak Ravanbakhsh,Chun-Liang Li,Peter Freeman,Barnabás Póczos +5 more
TL;DR: This work proposes a methodology based on Deep Generative Models to create complex models of galaxy morphologies that may meet the image simulation needs of upcoming surveys, and introduces GalSim-Hub, a community-driven repository of generative models, and a framework for incorporatingGenerative models within the GalSim image simulation software.
Proceedings ArticleDOI
Modeling Urbanization Patterns with Generative Adversarial Networks
TL;DR: A synthetic urban “universe” is generated that qualitatively reproduces the complex spatial organization observed in global urban patterns, while being able to quantitatively recover certain key high-level urban spatial metrics.
Journal ArticleDOI
Generative deep fields : arbitrarily sized, random synthetic astronomical images through deep learning
Michael J. Smith,James E. Geach +1 more
TL;DR: In this paper, a spatial-GAN was used to generate images resembling the iconic Hubble Space Telescope eXtreme deep field (XDF) in terms of abundance, morphology, magnitude distributions and colours.
References
More filters
Journal ArticleDOI
Bayesian-Based Iterative Method of Image Restoration
TL;DR: An iterative method of restoring degraded images was developed by treating images, point spread functions, and degraded images as probability-frequency functions and by applying Bayes’s theorem.
Proceedings Article
Generative adversarial text to image synthesis
TL;DR: In this article, a deep convolutional generative adversarial network (GAN) is used to generate plausible images of birds and flowers from detailed text descriptions, translating visual concepts from characters to pixels.
Posted Content
Generative Adversarial Text to Image Synthesis
TL;DR: A novel deep architecture and GAN formulation is developed to effectively bridge advances in text and image modeling, translating visual concepts from characters to pixels.
Posted Content
Euclid Definition Study Report
René J. Laureijs,Jérôme Amiaux,S. Arduini,J.-L. Auguères,Jarle Brinchmann,R. Cole,Mark Cropper,Christophe Dabin,Ludovic Duvet,Anne Ealet,Bianca Garilli,Philippe Gondoin,Luigi Guzzo,J. Hoar,Henk Hoekstra,Rory Holmes,Thomas D. Kitching,T. Maciaszek,Yannick Mellier,F. Pasian,Will J. Percival,Jason Rhodes,G. Saavedra Criado,Marc Sauvage,Roberto Scaramella,Luca Valenziano,Stephen J. Warren,Ralf Bender,Francisco J. Castander,Alessandro Cimatti,O. Le Fèvre,Hannu Kurki-Suonio,Michael Levi,P. B. Lilje,Georges Meylan,Robert C. Nichol,Kristian Pedersen,V. Popa,R. Rebolo Lopez,Hans-Walter Rix,H. J. A. Röttgering,Werner W. Zeilinger,Frank Grupp,P. Hudelot,Richard Massey,Massimo Meneghetti,Lance Miller,Stéphane Paltani,Stephane Paulin-Henriksson,Sandrine Pires,Curtis J. Saxton,Tim Schrabback,Gregor Seidel,Jeremy R. Walsh,Nabila Aghanim,Luca Amendola,James G. Bartlett,Carlo Baccigalupi,J.-P. Beaulieu,K. Benabed,Jean-Gabriel Cuby,David Elbaz,Pablo Fosalba,Giuseppe Gavazzi,Amina Helmi,I. M. Hook,Michael G. Irwin,Jean-Paul Kneib,Martin Kunz,Filippo Mannucci,Lauro Moscardini,Charling Tao,Romain Teyssier,Jochen Weller,G. Zamorani,M. R. Zapatero Osorio,Olivier Boulade,J. J. Foumond,A. Di Giorgio,P. Guttridge,A. James,A. James,M. Kemp,J. Martignac,A. P. Spencer,D. Walton,T. Blümchen,Carlotta Bonoli,Favio Bortoletto,C. Cerna,Leonardo Corcione,Christophe Fabron,Knud Jahnke,Sebastiano Ligori,F. Madrid,Laurent Martin,Gianluca Morgante,Tony Pamplona,Eric Prieto,Marco Riva,R. Toledo,M. Trifoglio,Filippo Maria Zerbi,F. B. Abdalla,Marian Douspis,C. Grenet,Stefano Borgani,R. J. Bouwens,Frederic Courbin,J.-M. Delouis,Pierre Dubath,Adriano Fontana,M. Frailis,Andrea Grazian,J. Koppenhöfer,O. Mansutti,M. Melchior,M. Mignoli,Joseph J. Mohr,C. Neissner,K. Noddle,M. Poncet,Marco Scodeggio,Santiago Serrano,Neville Shane,Jean-Luc Starck,Christian Surace,Andy Taylor,G. Verdoes-Kleijn,Claudio Vuerli,O. R. Williams,Andrea Zacchei,Bruno Altieri,I. Escudero Sanz,R. Kohley,T. Oosterbroek,Pierre Astier,D. J. Bacon,S. Bardelli,Carlton M. Baugh,F. Bellagamba,C. Benoist,Davide Bianchi,Andrea Biviano,E. Branchini,Carmelita Carbone,Vincenzo F. Cardone,Dylan N. Clements,Stephane Colombi,Christopher J. Conselice,Giovanni Cresci,Niall R. Deacon,James Dunlop,Cosimo Fedeli,F. Fontanot,P. Franzetti,Carlo Giocoli,Juan Garcia-Bellido,Jason Gow,Alan Heavens,P. Hewett,Catherine Heymans,Andrew D. Holland,Zhuoyi Huang,Olivier Ilbert,Benjamin Joachimi,E. Jennins,Eamonn Kerins,Alina Kiessling,Donnacha Kirk,Rubina Kotak,Oliver Krause,Ofer Lahav,F. van Leeuwen,Julien Lesgourgues,Marco Lombardi,Manuela Magliocchetti,Kate Maguire,Elisabetta Majerotto,R. Maoli,Federico Marulli,Sophie Maurogordato,H. J. McCracken,Ross J. McLure,Alessandro Melchiorri,Alex Merson,Michele Moresco,Mario Nonino,Peder Norberg,John A. Peacock,R. Pello,Matthew T. Penny,Valeria Pettorino,C. Di Porto,Lucia Pozzetti,Claudia Quercellini,Mario Radovich,Anais Rassat,Nicolas J.-H. Roche,Samuel Ronayette,Emanuel Rossetti,Barbara Sartoris,Barbara Sartoris,Peter Schneider,Elisabetta Semboloni,Stephen Serjeant,Fergus Simpson,Constantinos Skordis,G. Smadja,Stephen J. Smartt,P. Spano,S. Spiro,Mark Sullivan,Andre Tilquin,Roberto Trotta,Licia Verde,Y. Wang,G. Williger,G. Zhao,Julien Zoubian,E. Zucca +220 more
TL;DR: Euclid as mentioned in this paper is a space-based survey mission from the European Space Agency designed to understand the origin of the universe's accelerating expansion, using cosmological probes to investigate the nature of dark energy, dark matter and gravity by tracking their observational signatures.