Deeper detection limits in astronomical imaging using self-supervised spatiotemporal denoising

Abstract

The detection limit of astronomical imaging observations is limited by several noise sources. Some of that noise is correlated between neighbouring image pixels and exposures, so in principle could be learned and corrected. We present an astronomical self-supervised transformer-based denoising algorithm (ASTERIS), that integrates spatiotemporal information across multiple exposures. Benchmarking on mock data indicates that ASTERIS improves detection limits by 1.0 magnitude at 90% completeness and purity, while preserving the point spread function and photometric accuracy. Observational validation using data from the James Webb Space Telescope (JWST) and Subaru telescope identifies previously undetectable features, including low-surface-brightness galaxy structures and gravitationally-lensed arcs. Applied to deep JWST images, ASTERIS identifies three times more redshift > 9 galaxy candidates, with rest-frame ultraviolet luminosity 1.0 magnitude fainter, than previous methods.

Published in Science. This is the author's version of the work. It is posted here by permission of the AAAS for personal use, not for redistribution

References in corpus (36)

• Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising
• Galactic Stellar and Substellar Initial Mass Function
• Shot Noise in Mesoscopic Conductors
• EAZY: A Fast, Public Photometric Redshift Code
• Overview of the DESI Legacy Imaging Surveys
• A Comprehensive Study on Galaxies at z~9-16 Found in the Early JWST Data: UV Luminosity Functions and Cosmic Star-Formation History at the Pre-Reionization Epoch
• An updated analytic model for the attenuation by the intergalactic medium
• The Science Performance of JWST as Characterized in Commissioning
• New Determinations of the UV Luminosity Functions from z~9 to z~2 show a remarkable consistency with halo growth and a constant star formation efficiency
• Modelling and interpreting spectral energy distributions of galaxies with BEAGLE
• Cosmic Variance and Its Effect on the Luminosity Function Determination in Deep High z Surveys
• The spectral evolution of the first galaxies. I. James Webb Space Telescope detection limits and color criteria for population III galaxies
• Efficient Formation of Massive Galaxies at Cosmic Dawn by Feedback-Free Starbursts
• CEERS Epoch 1 NIRCam Imaging: Reduction Methods and Simulations Enabling Early JWST Science Results
• Spectroscopic confirmation of two luminous galaxies at $z\sim 14$
• Introducing the THESAN project: radiation-magnetohydrodynamic simulations of the epoch of reionization
• Confirmation and refutation of very luminous galaxies in the early universe
• The SDSS Coadd: 275 deg^2 of Deep SDSS Imaging on Stripe 82
• JADES NIRSpec Initial Data Release for the Hubble Ultra Deep Field: Redshifts and Line Fluxes of Distant Galaxies from the Deepest JWST Cycle 1 NIRSpec Multi-Object Spectroscopy
• The Universe at z>10: Predictions for JWST from the UniverseMachine DR1
• Pandeia: A Multi-mission Exposure Time Calculator for JWST and WFIRST
• Are the ultra-high-redshift galaxies at z > 10 surprising in the context of standard galaxy formation models?
• Morphologies of Galaxies at $z\gtrsim 9$ Uncovered by JWST/NIRCam Imaging: Cosmic Size Evolution and an Identification of an Extremely Compact Bright Galaxy at $z\sim 12$
• Early results from GLASS-JWST. V: the first rest-frame optical size-luminosity relation of galaxies at $z>7$
• Discovery of an Excess of Halpha Emitters around 4C 23.56 at z=2.48
• Learning to Denoise Astronomical Images with U-nets
• How dark the sky: the JWST backgrounds
• The Identification of a Dusty Multiarm Spiral Galaxy at $z=3.06$ with JWST and ALMA
• The SDSS Coadd: A Galaxy Photometric Redshift Catalog
• Inspiraling streams of enriched gas observed around a massive galaxy 11 billion years ago
• Photometric detection at $7.7\mu \mathrm{m}$ of a galaxy beyond redshift $14$ with JWST/MIRI
• Detection of Intracluster Globular Clusters in the First JWST Images of the Gravitational Lens Cluster SMACS J0723.3-7327 at z = 0.39
• Weak-signal extraction enabled by deep-neural-network denoising of diffraction data
• A Giant Disk Galaxy Two Billion Years After The Big Bang
• Astronomical image denoising by self-supervised deep learning and restoration processes
• Noise2Astro: Astronomical Image Denoising With Self-Supervised NeuralNetworks