Using deep learning for Detection, Instance Segmentation, and Classification on astronomical survey images.
Reference Paper: Merz et al. 2023
Corresponding Author: Grant Merz, University of Illinois at Urbana-Champaign
This is an updated repo of the original implementation (https://github.com/burke86/astro_rcnn)
DeepDISC is a deep learning framework for efficiently performing source detection, classification, and segmnetation (deblending) on astronomical images. We have built the code using detectron2 https://detectron2.readthedocs.io/en/latest/ for a modular design an access to state-of-the-art models.
- Create a conda environment. We recommend using python 3.9. You can use the environment.yml file provided and run
conda env create -f environment.yml
or create an environment from scratch and install by hand the packages listed in the environment.yml file
- Install deepdisc with
pip install deepdisc
You can also install by cloning this repo and runningpip install [e].[e] is optional and will install in editable mode. Use if you are going to change the source code
Usage:
demo_btk.ipynb
This notebook uses simulated images generated using the Blending Toolkit (Mendoza et al 2025). We used the CatSim (Connolly et al 2014) example catalog provided within the BTK to generate a set of training/test images, and then constructed per-image metadata the network needs to train. The notebook is largely for demo purposes, so does not include full training scheduling and optimizations.
The BTK simulated data can be downloaded here
Mendoza, Ismael, Andrii Torchylo, Thomas Sainrat, Axel Guinot, Alexandre Boucaud, Maxime Paillasa, Camille Avestruz, et al. 2025. “The Blending ToolKit: A Simulation Framework for Evaluation of Galaxy Detection and Deblending.” The Open Journal of Astrophysics 8 (February). https://doi.org/10.33232/001c.129699.
A.J. Connolly et al., An end-to-end simulation framework for the Large Synoptic Survey Telescope, in Proc. SPIE 9150 (2014) 14.