Skip to main content

"Unsupervised galaxy masking to remove background source detections"

Project description


DOI PyPI version Read the Docs

galmask is an open-source package written in Python that provides a simple way to remove unwanted background source detections from galaxy images. It builds on top of astropy and photutils astronomical Python libraries and the opencv and skimage image processing libraries.

The main requirements of galmask are:

  • astropy for handling FITS I/O and general-purpose astronomical routines.
  • photutils for photometry purposes and deblending detected sources.
  • opencv-python for connected-component analysis.
  • skimage for general image processing functionalities.


Via pip

galmask can be installed from PyPI via pip by running::

pip install galmask

Alternative method

galmask can also be installed by cloning the repository and doing a pip install in the project directory::

git clone
cd galmask
pip install .

It would be beneficial to create a python virtual environment and install the package within it, to prevent manipulating your global dependency versions.

Quick example

from import fits
from astropy.visualization import AsinhStretch, ImageNormalize, ZScaleInterval, LogStretch

import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1 import make_axes_locatable

# Import galmask
from galmask.galmask import galmask

def axes_colorbar(ax):
    divider = make_axes_locatable(ax)
    cax = divider.append_axes('bottom', size='5%', pad=0.3)
    return cax

filepath = 'example/gal1_G.fits'
image = fits.getdata(filepath)
npixels, nlevels, nsigma, contrast, min_distance, num_peaks, num_peaks_per_label, connectivity, remove_local_max = 5, 32, 2., 0.15, 1, 10, 3, 4, True  # Parameters for galmask
seg_image = None  # No segmentation map example

orig_segmap = fits.getdata('example/gal_seg1.fits')

galmasked, galsegmap = galmask(
    image, npixels, nlevels, nsigma, contrast, min_distance, num_peaks, num_peaks_per_label,
    connectivity=4, kernel=fits.getdata('kernel.fits'), seg_image=seg_image, mode="1",
    remove_local_max=True, deblend=True

# Plotting result.
fig, ax = plt.subplots(1, 4, figsize=(24, 6))

# For keeping original and final images on same scale.
vmin = min(image.min(), galmasked.min())
vmax = max(image.max(), galmasked.max())

# fig.suptitle(filepath)
norm1 = ImageNormalize(image, vmin=vmin, vmax=vmax, interval=ZScaleInterval(), stretch=LogStretch())
im0 = ax[0].imshow(image, norm=norm1, origin='lower', cmap='gray')
ax[0].set_title("Original image")
cax0 = axes_colorbar(ax[0])
fig.colorbar(im0, cax=cax0, orientation='horizontal')

im1 = ax[1].imshow(orig_segmap, origin='lower')
ax[1].set_title("Original segmentation map (photutils)")
cax1 = axes_colorbar(ax[1])
fig.colorbar(im1, cax=cax1, orientation='horizontal')

im2 = ax[2].imshow(galsegmap, origin='lower', cmap='gray')
ax[2].set_title("Final segmentation map (galmask)")
cax2 = axes_colorbar(ax[2])
fig.colorbar(im2, cax=cax2, orientation='horizontal')

norm2 = ImageNormalize(galmasked, vmin=vmin, vmax=vmax, interval=ZScaleInterval(), stretch=LogStretch())
im3 = ax[3].imshow(galmasked, norm=norm2, origin='lower', cmap='gray')
ax[3].set_title("Final image (galmask)")
cax3 = axes_colorbar(ax[3])
fig.colorbar(im3, cax=cax3, orientation='horizontal')



NOTE: orig_segmap is the original segmentation map - it is not returned by galmask. It is an intermediate result calculated inside galmask (if a pre-calculated segmentation map is not input). Here the original segmentation map was stored in a FITS file for demonstration purposes. So if you pass seg_image=None (as done in the above example) and would like to create such four-column plots, you would need to edit the source code of to save the internally calculated segmentation map in a FITS file.


The documentation is generated using the Sphinx documentation tool and hosted by Read the Docs. You can find the API reference and also some empirical tips to use galmask in the documentation.


For running the tests, you would need to install pytest. You can navigate to the tests/ directory and run:

pytest <name_of_file>


Contributions are welcome! Currently, there seem to be a few inefficient ways of handling things within galmask, and we would like you to contribute and improve the package!

Please let us know of any bugs/issues by opening an issue in the issue tracker.


License and copyright

galmask is licensed under the MIT License.

Copyright (c) 2022 Yash Gondhalekar

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

galmask-0.2.0.tar.gz (8.3 kB view hashes)

Uploaded source

Built Distribution

galmask-0.2.0-py3-none-any.whl (9.5 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor NVIDIA NVIDIA PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page