Probabilistic segmentation of overlapping galaxies for large cosmological surveys. (arXiv:2111.15455v1 [astro-ph.IM])
<a href="http://arxiv.org/find/astro-ph/1/au:+Hubert_B/0/1/0/all/0/1">Bretonnière Hubert</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Alexandre_B/0/1/0/all/0/1">Boucaud Alexandre</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Marc_H/0/1/0/all/0/1">Huertas-Company Marc</a>
Encoder-Decoder networks such as U-Nets have been applied successfully in a
wide range of computer vision tasks, especially for image segmentation of
different flavours across different fields. Nevertheless, most applications
lack of a satisfying quantification of the uncertainty of the prediction. Yet,
a well calibrated segmentation uncertainty can be a key element for scientific
applications such as precision cosmology. In this on-going work, we explore the
use of the probabilistic version of the U-Net, recently proposed by Kohl et al
(2018), and adapt it to automate the segmentation of galaxies for large
photometric surveys. We focus especially on the probabilistic segmentation of
overlapping galaxies, also known as blending. We show that, even when training
with a single ground truth per input sample, the model manages to properly
capture a pixel-wise uncertainty on the segmentation map. Such uncertainty can
then be propagated further down the analysis of the galaxy properties. To our
knowledge, this is the first time such an experiment is applied for galaxy
deblending in astrophysics.
Encoder-Decoder networks such as U-Nets have been applied successfully in a
wide range of computer vision tasks, especially for image segmentation of
different flavours across different fields. Nevertheless, most applications
lack of a satisfying quantification of the uncertainty of the prediction. Yet,
a well calibrated segmentation uncertainty can be a key element for scientific
applications such as precision cosmology. In this on-going work, we explore the
use of the probabilistic version of the U-Net, recently proposed by Kohl et al
(2018), and adapt it to automate the segmentation of galaxies for large
photometric surveys. We focus especially on the probabilistic segmentation of
overlapping galaxies, also known as blending. We show that, even when training
with a single ground truth per input sample, the model manages to properly
capture a pixel-wise uncertainty on the segmentation map. Such uncertainty can
then be propagated further down the analysis of the galaxy properties. To our
knowledge, this is the first time such an experiment is applied for galaxy
deblending in astrophysics.
http://arxiv.org/icons/sfx.gif
