Preprint — arXiv preprint arXiv:2510.22190 — 2025

RGC: a radio AGN classifier based on deep learning. I. A semi-supervised multiclass model for VLA images

MS Hossain, MSH Shahal, KMB Asad, P Saikia, A Khan, F Akter, A Ali, MA Amin, DP Guha, MOB Jihad, A Momen, S Sen, AKMM Rahman

arXiv preprint arXiv:2510.22190

Abstract

Context. Bent radio active galactic nuclei (RAGNs) – wide-angle tails (WATs) and narrow-angle tails (NATs) – trace dense environments in galaxy groups and clusters, yet no multi-class classifier simultaneously separates them from straight Fanaroff-Riley types (sFRI, sFRII) using visually inspected labels and unlabelled data.Aims. We release FIRST-2060, a four-class labelled dataset of 2060 RAGNs (sFRI, sFRII, WAT, NAT) constructed from three publicly available catalogues through multi-tier visual inspection, together with the semi-supervised RGC 1.0 model that leverages 20 000 unlabelled radio sources. We benchmarked RGC 1.0 against five supervised baselines spanning convolutional neural networks (CNNs) and vision transformers.Methods. FIRST-2060 is provided in two preprocessing variants: RL1, which retains spurious sources, and RL2, from which they are removed. The RGC model integrates the …

Cite

@article{u_35RYKgDlwC,
  title     = {RGC: a radio AGN classifier based on deep learning. I. A semi-supervised multiclass model for VLA images},
  author    = {MS Hossain and MSH Shahal and KMB Asad and P Saikia and A Khan and F Akter and A Ali and MA Amin and DP Guha and MOB Jihad and A Momen and S Sen and AKMM Rahman},
  journal = {arXiv preprint arXiv:2510.22190},
  year      = {2025}
}