Galaxy Morphological Classification Catalogue of the Dark Energy Survey Year 3 data with Convolutional Neural Networks. (arXiv:2107.10210v1 [astro-ph.GA])
<a href="http://arxiv.org/find/astro-ph/1/au:+Cheng_T/0/1/0/all/0/1">Ting-Yun Cheng</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Conselice_C/0/1/0/all/0/1">Christopher J. Conselice</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Aragon_Salamanca_A/0/1/0/all/0/1">Alfonso Arag&#xf3;n-Salamanca</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Aguena_M/0/1/0/all/0/1">M. Aguena</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Allam_S/0/1/0/all/0/1">S. Allam</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Andrade_Oliveira_F/0/1/0/all/0/1">F. Andrade-Oliveira</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Annis_J/0/1/0/all/0/1">J. Annis</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Bluck_A/0/1/0/all/0/1">A. F. L. Bluck</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Brooks_D/0/1/0/all/0/1">D. Brooks</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Burke_D/0/1/0/all/0/1">D. L. Burke</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Kind_M/0/1/0/all/0/1">M. Carrasco Kind</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Carretero_J/0/1/0/all/0/1">J. Carretero</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Choi_A/0/1/0/all/0/1">A. Choi</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Costanzi_M/0/1/0/all/0/1">M. Costanzi</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Costa_L/0/1/0/all/0/1">L. N. da Costa</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Pereira_M/0/1/0/all/0/1">M. E. S. Pereira</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Vicente_J/0/1/0/all/0/1">J. De Vicente</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Diehl_H/0/1/0/all/0/1">H. T. Diehl</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Drlica_Wagner_A/0/1/0/all/0/1">A. Drlica-Wagner</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Eckert_K/0/1/0/all/0/1">K. Eckert</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Everett_S/0/1/0/all/0/1">S. Everett</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Evrard_A/0/1/0/all/0/1">A. E. Evrard</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Ferrero_I/0/1/0/all/0/1">I. Ferrero</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Fosalba_P/0/1/0/all/0/1">P. Fosalba</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Frieman_J/0/1/0/all/0/1">J. Frieman</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Garcia_Bellido_J/0/1/0/all/0/1">J. Garc&#xed;a-Bellido</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Gerdes_D/0/1/0/all/0/1">D. W. Gerdes</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Giannantonio_T/0/1/0/all/0/1">T. Giannantonio</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Gruen_D/0/1/0/all/0/1">D. Gruen</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Gruendl_R/0/1/0/all/0/1">R. A. Gruendl</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Gschwend_J/0/1/0/all/0/1">J. Gschwend</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Gutierrez_G/0/1/0/all/0/1">G. Gutierrez</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Hinton_S/0/1/0/all/0/1">S. R. Hinton</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Hollowood_D/0/1/0/all/0/1">D. L. Hollowood</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Honscheid_K/0/1/0/all/0/1">K. Honscheid</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+James_D/0/1/0/all/0/1">D. J. James</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Krause_E/0/1/0/all/0/1">E. Krause</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Kuehn_K/0/1/0/all/0/1">K. Kuehn</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Kuropatkin_N/0/1/0/all/0/1">N. Kuropatkin</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Lahav_O/0/1/0/all/0/1">O. Lahav</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Maia_M/0/1/0/all/0/1">M. A. G. Maia</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+March_M/0/1/0/all/0/1">M. March</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Menanteau_F/0/1/0/all/0/1">F. Menanteau</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Miquel_R/0/1/0/all/0/1">R. Miquel</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Morgan_R/0/1/0/all/0/1">R. Morgan</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Paz_Chinchon_F/0/1/0/all/0/1">F. Paz-Chinch&#xf3;n</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Pieres_A/0/1/0/all/0/1">A. Pieres</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Malagon_A/0/1/0/all/0/1">A. A. Plazas Malag&#xf3;n</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Roodman_A/0/1/0/all/0/1">A. Roodman</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Sanchez_E/0/1/0/all/0/1">E. Sanchez</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Scarpine_V/0/1/0/all/0/1">V. Scarpine</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Serrano_S/0/1/0/all/0/1">S. Serrano</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Sevilla_Noarbe_I/0/1/0/all/0/1">I. Sevilla-Noarbe</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Smith_M/0/1/0/all/0/1">M. Smith</a>, <a href="http://arxiv.org/find/astro-ph/1/au:+Soares_Santos_M/0/1/0/all/0/1">M. Soares-Santos</a>, et al. (5 additional authors not shown)

We present in this paper one of the largest galaxy morphological
classification catalogues to date, including over 20 million of galaxies, using
the Dark Energy Survey (DES) Year 3 data based on Convolutional Neural Networks
(CNN). Monochromatic $i$-band DES images with linear, logarithmic, and gradient
scales, matched with debiased visual classifications from the Galaxy Zoo 1
(GZ1) catalogue, are used to train our CNN models. With a training set
including bright galaxies ($16le{i}<18$) at low redshift ($z<0.25$), we
furthermore investigate the limit of the accuracy of our predictions applied to
galaxies at fainter magnitude and at higher redshifts. Our final catalogue
covers magnitudes $16le{i}<21$, and redshifts $z<1.0$, and provides predicted
probabilities to two galaxy types — Ellipticals and Spirals (disk galaxies).
Our CNN classifications reveal an accuracy of over 99% for bright galaxies
when comparing with the GZ1 classifications ($i<18$). For fainter galaxies, the
visual classification carried out by three of the co-authors shows that the CNN
classifier correctly categorises disky galaxies with rounder and blurred
features, which humans often incorrectly visually classify as Ellipticals. As a
part of the validation, we carry out one of the largest examination of
non-parametric methods, including $sim$100,000 galaxies with the same coverage
of magnitude and redshift as the training set from our catalogue. We find that
the Gini coefficient is the best single parameter discriminator between
Ellipticals and Spirals for this data set.

We present in this paper one of the largest galaxy morphological
classification catalogues to date, including over 20 million of galaxies, using
the Dark Energy Survey (DES) Year 3 data based on Convolutional Neural Networks
(CNN). Monochromatic $i$-band DES images with linear, logarithmic, and gradient
scales, matched with debiased visual classifications from the Galaxy Zoo 1
(GZ1) catalogue, are used to train our CNN models. With a training set
including bright galaxies ($16le{i}<18$) at low redshift ($z<0.25$), we
furthermore investigate the limit of the accuracy of our predictions applied to
galaxies at fainter magnitude and at higher redshifts. Our final catalogue
covers magnitudes $16le{i}<21$, and redshifts $z<1.0$, and provides predicted
probabilities to two galaxy types — Ellipticals and Spirals (disk galaxies).
Our CNN classifications reveal an accuracy of over 99% for bright galaxies
when comparing with the GZ1 classifications ($i<18$). For fainter galaxies, the
visual classification carried out by three of the co-authors shows that the CNN
classifier correctly categorises disky galaxies with rounder and blurred
features, which humans often incorrectly visually classify as Ellipticals. As a
part of the validation, we carry out one of the largest examination of
non-parametric methods, including $sim$100,000 galaxies with the same coverage
of magnitude and redshift as the training set from our catalogue. We find that
the Gini coefficient is the best single parameter discriminator between
Ellipticals and Spirals for this data set.

http://arxiv.org/icons/sfx.gif