Skip to content

Latest commit

 

History

History
28 lines (20 loc) · 911 Bytes

README.md

File metadata and controls

28 lines (20 loc) · 911 Bytes
Raw

Star-galaxy Classification Using Deep Convolutional Neural Networks

  • Edward J Kim
  • Robert J Brunner

Paper

The LaTex source files are in the paper directory.

IPython/Jupyter Notebooks

How to generate SDSS cutout images: This notebook demonstrates how to generate cutout images from the SDSS database. Note: For demonstration and testing purposes only. For creating a training set, use scripts/fetch_sdss.py.

Training ConvNets: This notebook demonstrates the ConvNet architecture with a sample training set of 100 images. For demonstration and testing purposes only. In the paper, we used a much larger training set. See scripts/train_cnn.py.

Dockerfile

The Dockerfile for the Jupyter notebook server that has been used to run the notebooks is provided in the docker directory.