Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Mapping Galaxy Images Across Ultraviolet, Visible and Infrared Bands Using Generative Deep Learning

Youssef Zaazou, Alex Bihlo, Terrence S. Tricco

TL;DR

The paper tackles the problem of translating galaxy images across photometric bands to enable robust multi-band analyses and efficient survey planning. It introduces a fully supervised image-to-image generator capable of both band interpolation and extrapolation, trained on Illustris mock observations and validated with DECaLS data. The approach achieves high fidelity according to general metrics (MAE, SSIM, PSNR) and preserves galaxy morphology as shown by GINI and M20, with $W_1$ distances confirming distributional alignment. This work enables enhanced exploration of multi-band information in incomplete data regions and supports optimized mission planning and targeted follow-ups, with code accessible for reproducibility.

Abstract

We demonstrate that generative deep learning can translate galaxy observations across ultraviolet, visible, and infrared photometric bands. Leveraging mock observations from the Illustris simulations, we develop and validate a supervised image-to-image model capable of performing both band interpolation and extrapolation. The resulting trained models exhibit high fidelity in generating outputs, as verified by both general image comparison metrics (MAE, SSIM, PSNR) and specialized astronomical metrics (GINI coefficient, M20). Moreover, we show that our model can be used to predict real-world observations, using data from the DECaLS survey as a case study. These findings highlight the potential of generative learning to augment astronomical datasets, enabling efficient exploration of multi-band information in regions where observations are incomplete. This work opens new pathways for optimizing mission planning, guiding high-resolution follow-ups, and enhancing our understanding of galaxy morphology and evolution.

Mapping Galaxy Images Across Ultraviolet, Visible and Infrared Bands Using Generative Deep Learning

TL;DR

The paper tackles the problem of translating galaxy images across photometric bands to enable robust multi-band analyses and efficient survey planning. It introduces a fully supervised image-to-image generator capable of both band interpolation and extrapolation, trained on Illustris mock observations and validated with DECaLS data. The approach achieves high fidelity according to general metrics (MAE, SSIM, PSNR) and preserves galaxy morphology as shown by GINI and M20, with distances confirming distributional alignment. This work enables enhanced exploration of multi-band information in incomplete data regions and supports optimized mission planning and targeted follow-ups, with code accessible for reproducibility.

Abstract

We demonstrate that generative deep learning can translate galaxy observations across ultraviolet, visible, and infrared photometric bands. Leveraging mock observations from the Illustris simulations, we develop and validate a supervised image-to-image model capable of performing both band interpolation and extrapolation. The resulting trained models exhibit high fidelity in generating outputs, as verified by both general image comparison metrics (MAE, SSIM, PSNR) and specialized astronomical metrics (GINI coefficient, M20). Moreover, we show that our model can be used to predict real-world observations, using data from the DECaLS survey as a case study. These findings highlight the potential of generative learning to augment astronomical datasets, enabling efficient exploration of multi-band information in regions where observations are incomplete. This work opens new pathways for optimizing mission planning, guiding high-resolution follow-ups, and enhancing our understanding of galaxy morphology and evolution.
Paper Structure (14 sections, 2 equations, 1 figure, 1 table)

This paper contains 14 sections, 2 equations, 1 figure, 1 table.

Table of Contents

  1. Introduction
  2. Image-To-Image Translation
  3. Deep Learning Applications in Astronomy
  4. Our research
  5. Proposed Approach
  6. Dataset
  7. Illustris Dataset
  8. Galaxy 10 Decals Dataset
  9. Metrics
  10. Inference Techniques
  11. Generator Training
  12. Generator Architecture
  13. Results
  14. Interpolation

Figures (1)

  • Figure 1: The results for our three trained interpolation models; each column is reserved for a single galaxy. The $NUV$ inputs and the $K$ inputs ($X_{NUV}$ and $X_{K}$) are delineated by the $G$, $R$, and $Z$ groupings (one for each model).The first row of each grouping shows the ground truth labels, the second shows the respective generated images while the last row is the residual taken to be the ground truth label minus the model output.