Radio-Interferometric Image Reconstruction with Denoising Diffusion Restoration Models

M. Morales, 1 E. Tolley, 1 R. Poitevineau 1 1 Institute of Physics, Laboratory of Astrophysics, École Polytechnique Fédérale de Lausanne (EPFL), 1290 Sauverny, Switzerland · 2026

Read Radio-Interferometric Image Reconstruction with Denoising Diffusion Restoration Models by M. Morales, 1 E. Tolley, 1 R. Poitevineau 1 1 Institute of Physics, Laboratory of Astrophysics, École Polytechnique Fédérale de Lausanne (EPFL), 1290 Sauverny, Switzerland (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-15844.txt.

Description

M. Morales, E. Tolley, and R. Poitevineau rebuild radio sky images with AI in a 2026 arXiv paper. Radio interferometers sample the sky incompletely. They measure Fourier components, not full images. Reconstructing a picture from partial data is a key challenge in radio astronomy. The authors use denoising diffusion restoration models. Their method learns a data-driven prior for the radio sky. That prior fills gaps in a principled way.

They first train a diffusion model on real observations. The training set is radio galaxy data from the VLA FIRST survey. A learned prior knows what radio galaxies tend to look like. Reconstructions can then favor plausible skies over noise. For upcoming large surveys, fast and faithful imaging is a core need.

Catalog record

CREATORM. Morales, 1 E. Tolley, 1 R. Poitevineau 1 1 Institute of Physics, Laboratory of Astrophysics, École Polytechnique Fédérale de Lausanne (EPFL), 1290 Sauverny, Switzerland
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE27 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-15844

Browse the full library · About the Tower of Babel