A Value-added Physical Properties Catalog for Low-redshift Galaxies from DESI Legacy Imaging Surveys DR10
Shirui Wei, Changhua Li, Yanxia Zhang, Chenzhou Cui, Jinghang Shi, Wujun Shao, Zihan Kang, Yongheng Zhao, Maoyuan Huang

TL;DR
This paper presents a deep learning-based catalog estimating star formation rate, stellar mass, and metallicity for over 500 million low-redshift galaxies from DESI imaging surveys, enabling large-scale population studies.
Contribution
Introduces a multimodal deep learning model combining imaging and catalog data to efficiently estimate galaxy properties for large photometric datasets.
Findings
Generated a catalog for ~547 million galaxies with property estimates.
Validated the model against independent measurements and scaling relations.
Provided the first homogeneous photometry-based property catalog for DESI LS DR10.
Abstract
Galaxy physical properties-such as star formation rate (SFR), stellar mass, and gas-phase metallicity-are essential for population studies and evolutionary analyses. Deriving these quantities for billions of galaxies in modern imaging surveys presents significant challenges due to limited spectroscopy and the computational costs associated with traditional spectral energy distribution fitting. As a result, many galaxies in large photometric surveys still lack homogeneous property estimates. This study introduces a multimodal deep learning model that integrates optical imaging with photometric catalog features to estimate SFR, stellar mass, and oxygen abundance in low-redshift galaxies. The model incorporates a ResNet-based convolutional neural network to extract spatial information from multiband images and a multilayer perceptron that processes catalog-level photometric features,…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
