Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Galaxy Light profile neural Networks (GaLNets). II. Bulge-Disc decomposition in optical space-based observations

Chen Qiu, Nicola R. Napolitano, Rui Li, Yuedong Fang, Crescenzo Tortora, Shiyin Shen, Luis C. Ho, Weipeng Lin, Leyao Wei, Ran Li, Zuhui Fan, Yang Wang, Guoliang Li, Hu Zhan, Dezi Liu

TL;DR

The paper advances automated bulge-disk decomposition for large, high-redshift galaxy samples by extending GaLNet to PSF-convolved 2-Sérsic fits in space-based CSST data. It presents GaLNet-BD, a CNN-based regressor that predicts BD structure from two inputs (galaxy image and PSF) and is trained on realistic CSST mock images built from CosmoDC2-derived parameters. Results show GaLNet-BD achieving high accuracy for most BD parameters down to r ≈ 23.5 at z ≈ 1, with particularly strong performance on magnitudes and bulge-to-total estimates; complementary tests with 1-Sérsic GaLNet indicate viable analyses with half-depth CSST data up to r ≈ 24 and z ≈ 1.7. These findings demonstrate the potential for rapid, scalable extraction of galaxy structural parameters across billions of galaxies in upcoming space surveys, enabling robust morphological and evolutionary studies.

Abstract

Bulge-disk (B-D) decomposition is an effective diagnostic to characterize the galaxy morphology and understand its evolution across time. So far, high-quality data have allowed detailed B-D decomposition to redshift below 0.5, with limited excursions over small volumes at higher redshifts. Next-generation large sky space surveys in optical, e.g. from the China Space Station Telescope (CSST), and near-infrared, e.g. from the space EUCLID mission, will produce a gigantic leap in these studies as they will provide deep, high-quality photometric images over more than 15000 deg2 of the sky, including billions of galaxies. Here, we extend the use of the Galaxy Light profile neural Network (GaLNet) to predict 2-Sérsic model parameters, specifically from CSST data. We simulate point-spread function (PSF) convolved galaxies, with realistic B-D parameter distributions, on CSST mock observations to train the new GaLNet and predict the structural parameters (e.g. magnitude, effective radius, Sersic index, axis ratio, etc.) of both bulge and disk components. We find that the GaLNet can achieve very good accuracy for most of the B-D parameters down to an $r$-band magnitude of 23.5 and redshift $\sim$1. The best accuracy is obtained for magnitudes, implying accurate bulge-to-total (B/T) estimates. To further forecast the CSST performances, we also discuss the results of the 1-Sérsic GaLNet and show that CSST half-depth data will allow us to derive accurate 1-component models up to $r\sim$24 and redshift z$\sim$1.7.

Galaxy Light profile neural Networks (GaLNets). II. Bulge-Disc decomposition in optical space-based observations

TL;DR

The paper advances automated bulge-disk decomposition for large, high-redshift galaxy samples by extending GaLNet to PSF-convolved 2-Sérsic fits in space-based CSST data. It presents GaLNet-BD, a CNN-based regressor that predicts BD structure from two inputs (galaxy image and PSF) and is trained on realistic CSST mock images built from CosmoDC2-derived parameters. Results show GaLNet-BD achieving high accuracy for most BD parameters down to r ≈ 23.5 at z ≈ 1, with particularly strong performance on magnitudes and bulge-to-total estimates; complementary tests with 1-Sérsic GaLNet indicate viable analyses with half-depth CSST data up to r ≈ 24 and z ≈ 1.7. These findings demonstrate the potential for rapid, scalable extraction of galaxy structural parameters across billions of galaxies in upcoming space surveys, enabling robust morphological and evolutionary studies.

Abstract

Bulge-disk (B-D) decomposition is an effective diagnostic to characterize the galaxy morphology and understand its evolution across time. So far, high-quality data have allowed detailed B-D decomposition to redshift below 0.5, with limited excursions over small volumes at higher redshifts. Next-generation large sky space surveys in optical, e.g. from the China Space Station Telescope (CSST), and near-infrared, e.g. from the space EUCLID mission, will produce a gigantic leap in these studies as they will provide deep, high-quality photometric images over more than 15000 deg2 of the sky, including billions of galaxies. Here, we extend the use of the Galaxy Light profile neural Network (GaLNet) to predict 2-Sérsic model parameters, specifically from CSST data. We simulate point-spread function (PSF) convolved galaxies, with realistic B-D parameter distributions, on CSST mock observations to train the new GaLNet and predict the structural parameters (e.g. magnitude, effective radius, Sersic index, axis ratio, etc.) of both bulge and disk components. We find that the GaLNet can achieve very good accuracy for most of the B-D parameters down to an -band magnitude of 23.5 and redshift 1. The best accuracy is obtained for magnitudes, implying accurate bulge-to-total (B/T) estimates. To further forecast the CSST performances, we also discuss the results of the 1-Sérsic GaLNet and show that CSST half-depth data will allow us to derive accurate 1-component models up to 24 and redshift z1.7.
Paper Structure (7 sections, 7 equations, 4 figures)

This paper contains 7 sections, 7 equations, 4 figures.

Table of Contents

  1. Introduction
  2. Methods and Data
  3. A GaLNet for Bulge-Disk decomposition: GaLNet-BD
  4. The CNN Architecture
  5. Data: Mock CSST galaxy cutouts
  6. CSST simulations
  7. Simulating realistic B-D galaxy images

Figures (4)

  • Figure 1: 1D normalized Sérsic surface brightness profiles in linear (top) and "log" scale (bottom) as a function of the radius in units of $R_{\rm e}$, for different $n$-indices ($n=0.5,1,2,4,5,7$, from blue to purple).
  • Figure 2: The structure of the GaLNet-BD used in this work. The networks are fed by both galaxy images and the corresponding “local” simulated PSFs, and the outputs are the 10 parameters of the two Sérsic profiles. 6 layers are used for the galaxy branch and 4 layers for the PSF branch. After the Concatenation layer, we add 3 fully connected layers to extract further features of the combined galaxies and PSFs.
  • Figure 3: Structural and positional parameters are used in fitting Sérsic model for both bulges and disks. To make realistic simulations, we combined 2-Sérsic model convolved with PSF and CSST background value (see §\ref{['sec:sim_gal']}).
  • Figure 4: A set of reduced $r$-band CSST simulated images (first row) and the same images with simulated galaxies according to the procedure illustrated in §\ref{['sec:sim_gal']} (second row).