Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Estimating Flux Densities of Diffuse Cosmological Radio Sources Exploiting Vision Transformers

Nicoletta Sanvitale, Claudio Gheller, Franco Vazza, Federica Govoni, Matteo Murgia, Valentina Vacca

TL;DR

This work tackles the challenge of measuring ultra-faint diffuse radio emission from the Warm–Hot Intergalactic Medium in large radio surveys. It introduces TUNA, a Vision‑Transformer–based network adapted from segmentation to flux regression, trained on LOFAR‑like mock observations generated from cosmological MHD simulations. The study demonstrates that TUNA can accurately recover low surface brightness structures with mild smoothing and quantify biases, and it validates the approach on real LOFAR data by recovering a ridge in the A399–A401 field that is not visible in high‑resolution inputs and is consistent with a low‑resolution tapered map. The results suggest that automated, quantitative surface brightness estimates from transformer‑based architectures can enable scalable analyses for upcoming facilities like the SKA, provided calibration and generalization are further improved.

Abstract

We present TUNA, a Vision-Transformer based network adapted from segmentation to flux regression for faint, diffuse radio emission. Trained on LOFAR-like mock observations derived from cosmological simulations, TUNA accurately reconstructs low surface-brightness structures, with only mild smoothing and small brightness-dependent biases. Applied to LOFAR data of the A399 - A401 galaxy cluster system, it recovers the ridge not identifiable in the high resolution observation and matches the low resolution tapered map. These results indicate how TUNA can deliver automated, quantitative surface brightness estimates for diffuse extragalactic sources, enabling scalable analyses for upcoming surveys.

Estimating Flux Densities of Diffuse Cosmological Radio Sources Exploiting Vision Transformers

TL;DR

This work tackles the challenge of measuring ultra-faint diffuse radio emission from the Warm–Hot Intergalactic Medium in large radio surveys. It introduces TUNA, a Vision‑Transformer–based network adapted from segmentation to flux regression, trained on LOFAR‑like mock observations generated from cosmological MHD simulations. The study demonstrates that TUNA can accurately recover low surface brightness structures with mild smoothing and quantify biases, and it validates the approach on real LOFAR data by recovering a ridge in the A399–A401 field that is not visible in high‑resolution inputs and is consistent with a low‑resolution tapered map. The results suggest that automated, quantitative surface brightness estimates from transformer‑based architectures can enable scalable analyses for upcoming facilities like the SKA, provided calibration and generalization are further improved.

Abstract

We present TUNA, a Vision-Transformer based network adapted from segmentation to flux regression for faint, diffuse radio emission. Trained on LOFAR-like mock observations derived from cosmological simulations, TUNA accurately reconstructs low surface-brightness structures, with only mild smoothing and small brightness-dependent biases. Applied to LOFAR data of the A399 - A401 galaxy cluster system, it recovers the ridge not identifiable in the high resolution observation and matches the low resolution tapered map. These results indicate how TUNA can deliver automated, quantitative surface brightness estimates for diffuse extragalactic sources, enabling scalable analyses for upcoming surveys.
Paper Structure (3 sections, 4 figures)

This paper contains 3 sections, 4 figures.

Table of Contents

  1. Introduction
  2. Methodology
  3. Results and Discussion

Figures (4)

  • Figure 1: Example of a $1.1^\circ \times 1.1^\circ$ mock image used as input for the network (left) and the corresponding prediction from TUNA (center). The sky map (our ground truth) is shown on the right.
  • Figure 2: Distribution of surface brightness from the test and prediction data with surface brightness higher than $10^{-8}$Jy/arcsec$^2$.
  • Figure 3: Predicted surface brightness distribution for the A399-A401 system (left panel) compared with the input image (centre) and the $80$ arcsec tapered image (right). All panels are thresholded at 1$\sigma$, where $\sigma$ is the rms noise measured in each image ($\sigma_\mathrm{Inference} = 1.22\times 10^{-7}$ Jy/arcsec$^2$, $\sigma_\mathrm{Input} = 2.65\times 10^{-6}$ Jy/arcsec$^2$, $\sigma_\mathrm{Tapered} = 1.38\times 10^{-7}$ Jy/arcsec$^2$). The box and cyan line indicate the regions where the flux density and the surface brightness have been evaluated.
  • Figure 4: Zoom-in on the 1D surface brightness profile across the ridge region. The plot shows the predicted (blue), input (red), and tapered (green) surface brightness curves, each thresholded at 0.5$\sigma$. Vertical dashed lines indicate the boundaries of the yellow box used for flux integration.