Can AI Dream of Unseen Galaxies? Conditional Diffusion Model for Galaxy Morphology Augmentation
Open Access
- 13 January 2026
- journal article
- Published by American Astronomical Society in The Astrophysical Journal Supplement Series
- Vol. 282 (2) , 25
Abstract
Observational astronomy relies on visual feature identification to detect critical astrophysical phenomena. While machine learning (ML) increasingly automates this process, models often struggle with generalization in large-scale surveys due to the limited representativeness of labeled datasets—whether from simulations or human annotation—a challenge pronounced for rare yet scientifically valuable objects. To address this, we propose a conditional diffusion model to synthesize realistic galaxy images for augmenting ML training data (hereafter GalaxySD). Leveraging the Galaxy Zoo 2 (GZ2) dataset, which contains visual feature–galaxy image pairs from volunteer annotation, we demonstrate that GalaxySD generates diverse, high-fidelity galaxy images that closely adhere to the specified morphological feature conditions. Moreover, this model enables generative extrapolation to project well-annotated data into unseen domains and advance rare object detection. Integrating synthesized images into ML pipelines improves performance in standard morphology classification, boosting completeness and purity by up to 30% across key metrics. For rare object detection, using early-type galaxies with prominent dust lane features (∼0.1% in the GZ2 dataset) as a test case, our approach doubled the number of detected instances—from 352 to 872—compared to previous studies based on visual inspection. This study highlights the power of generative models to bridge gaps between scarce labeled data and the vast, uncharted parameter space of observational astronomy and sheds insight for future astrophysical foundation model developments. Our project homepage is available at https://galaxysd-webpage.streamlit.app/ .Keywords
All Related Versions
Funding Information
- MOST ∣ National Natural Science Foundation of China (12273015)
- MOST ∣ National Key Research and Development Program of China (2023YFA1605600)
- Tsinghua University (20223080023)
This publication has 115 references indexed in Scilit:
- Astronomical image denoising by self-supervised deep learning and restoration processesNature Astronomy, 2025
- Evaluation metrics for galaxy image generatorsAstronomy and Computing, 2023
- DeepGhostBusters: Using Mask R-CNN to detect and mask ghosting and scattered-light artifacts from optical survey imagesAstronomy and Computing, 2022
- A cold, massive, rotating disk galaxy 1.5 billion years after the Big BangNature, 2020
- Interacting galaxies and mergersNature Astronomy, 2019
- The strong gravitational lens finding challengeAstronomy & Astrophysics, 2019
- Systematic search for tidal features around nearby galaxiesAstronomy & Astrophysics, 2018
- GalSim: The modular galaxy image simulation toolkitAstronomy and Computing, 2015
- Analysis of galaxy spectral energy distributions from far-UV to far-IR with CIGALE: studying a SINGS test sampleAstronomy & Astrophysics, 2009
- Stellar population synthesis at the resolution of 2003Monthly Notices of the Royal Astronomical Society, 2003