Deep Learning for space-variant deconvolution in galaxy surveys. (arXiv:1911.00443v1 [astro-ph.IM])
<a href="http://arxiv.org/find/astro-ph/1/au:+Sureau_F/0/1/0/all/0/1">Florent Sureau</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Lechat_A/0/1/0/all/0/1">Alexis Lechat</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Starck_J/0/1/0/all/0/1">Jean-Luc Starck</a>
Deconvolution of large survey images with millions of galaxies requires to
develop a new generation of methods which can take into account a space variant
Point Spread Function and have to be at the same time accurate and fast. We
investigate in this paper how Deep Learning could be used to perform this task.
We employ a U-NET Deep Neural Network architecture to learn in a supervised
setting parameters adapted for galaxy image processing and study two strategies
for deconvolution. The first approach is a post-processing of a mere Tikhonov
deconvolution with closed form solution and the second one is an iterative
deconvolution framework based on the Alternating Direction Method of
Multipliers (ADMM). Our numerical results based on GREAT3 simulations with
realistic galaxy images and PSFs show that our two approaches outperforms
standard techniques based on convex optimization, whether assessed in galaxy
image reconstruction or shape recovery. The approach based on Tikhonov
deconvolution leads to the most accurate results except for ellipticity errors
at high signal to noise ratio where the ADMM approach performs slightly better,
is also more computation-time efficient to process a large number of galaxies,
and is therefore recommended in this scenario.
Deconvolution of large survey images with millions of galaxies requires to
develop a new generation of methods which can take into account a space variant
Point Spread Function and have to be at the same time accurate and fast. We
investigate in this paper how Deep Learning could be used to perform this task.
We employ a U-NET Deep Neural Network architecture to learn in a supervised
setting parameters adapted for galaxy image processing and study two strategies
for deconvolution. The first approach is a post-processing of a mere Tikhonov
deconvolution with closed form solution and the second one is an iterative
deconvolution framework based on the Alternating Direction Method of
Multipliers (ADMM). Our numerical results based on GREAT3 simulations with
realistic galaxy images and PSFs show that our two approaches outperforms
standard techniques based on convex optimization, whether assessed in galaxy
image reconstruction or shape recovery. The approach based on Tikhonov
deconvolution leads to the most accurate results except for ellipticity errors
at high signal to noise ratio where the ADMM approach performs slightly better,
is also more computation-time efficient to process a large number of galaxies,
and is therefore recommended in this scenario.
http://arxiv.org/icons/sfx.gif
