J/MNRAS/476/3661 Morphology of SDSS galaxies (Dominguez Sanchez+, 2018)
Improving galaxy morphologies for SDSS with deep learning.
Dominguez Sanchez H., Huertas-Company M., Bernardi M., Tuccillo D.,
Fischer J.L.
<Mon. Not. R. Astron. Soc., 476, 3661-3676 (2018)>
=2018MNRAS.476.3661D 2018MNRAS.476.3661D (SIMBAD/NED BibCode)
ADC_Keywords: Galaxy catalogs ; Morphology
Keywords: methods: observational - catalogues - galaxies: structure
Abstract:
We present a morphological catalogue for ∼670000 galaxies in the
Sloan Digital Sky Survey in two flavours: T-type, related to the
Hubble sequence, and Galaxy Zoo 2 (GZ2 hereafter) classification
scheme. By combining accurate existing visual classification
catalogues with machine learning, we provide the largest and most
accurate morphological catalogue up to date. The classifications are
obtained with Deep Learning algorithms using Convolutional Neural
Networks (CNNs). We use two visual classification catalogues, GZ2 and
Nair & Abraham (2010ApJS..186..427N 2010ApJS..186..427N, Cat. J/ApJS/186/427), for
training CNNs with colour images in order to obtain T-types and a
series of GZ2 type questions (disc/features, edge-on galaxies, bar
signature, bulge prominence, roundness, and mergers). We also provide
an additional probability enabling a separation between pure
elliptical (E) from S0, where the T-type model is not so efficient.
For the T-type, our results show smaller offset and scatter than
previous models trained with support vector machines. For the GZ2 type
questions, our models have large accuracy (>97 per cent), precision
and recall values (>90 per cent), when applied to a test sample with
the same characteristics as the one used for training. The catalogue
is publicly released with the paper.
Description:
We present a morphological catalogue for a sample of ∼670000 galaxies
from the SDSS DR7 corresponding to the sample analysed by Meert et al.
(2015MNRAS.446.3943M 2015MNRAS.446.3943M, Cat. J/MNRAS/446/3943; 2016MNRAS.455.2440M 2016MNRAS.455.2440M, Cat.
J/MNRAS/455/2440). The morphological classifications are obtained with
Deep Learning algorithms using CNNs, and the models are trained with
the best available visual classification catalogues (Nair & Abraham,
2010ApJS..186..427N 2010ApJS..186..427N, Cat. J/ApJS/186/427; Willett et al.
2013MNRAS.435.2835W 2013MNRAS.435.2835W, Cat. J/MNRAS/435/2835).
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
catalog.dat 131 670722 *Catalogue released with this paper
--------------------------------------------------------------------------------
Note on catalog.dat: The catalogue contains 670722 rows, each corresponding to a
galaxy from the M15 sample. The last column of this table indicates which
catalogue has been used for training each model.
--------------------------------------------------------------------------------
See also:
J/ApJS/186/427 : Detailed morphology of SDSS galaxies (Nair+, 2010)
J/MNRAS/435/2835 : Morphological types from Galaxy Zoo 2 (Willett+, 2013)
J/MNRAS/446/3943 : Galaxies 2D phot. decompositions in SDSS-DR7 (Meert+, 2015)
J/MNRAS/455/2440 : Gal. 2D phot. decompositions in r, g+i bands (Meert+, 2016)
Byte-by-byte Description of file: catalog.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 19 I19 --- objID SDSS DR7 objID (dr7objid)
21- 26 I6 --- M15 Meert et al. (2015, Cat. J/MNRAS/446/3943)
ID (galcount)
28- 38 E11.6 --- Pdisk Probability of showing features/disc
(vs. being smooth) (P_disk)
40- 50 E11.6 --- Pedgeon Probability of being edge on (Pedgeon)
52- 62 E11.6 --- PbarGZ2 Probability of having bar signature
(trained with Galaxy Zoo 2 catalogue)
(PbarGZ2)
64- 74 E11.6 --- PbarNair10 Probability of having bar signature
(trained with Nair et al., 2010,
J/ApJS/186/427 catalogue) (PbarNair10)
76- 86 E11.6 --- Pmerg Probability of being merger/projected pairs
(P_merg) (1)
88- 98 E11.6 --- Pbulge Probability of having a dominant/obvious
bulge (vs. no bulge) (P_bulge)
100-110 E11.6 --- Pcigar Probability of having cigar shape
(vs. round shape) (P_cigar)
112-119 F8.5 --- TType T-Type (TType) (2)
121-131 E11.6 --- PS0 Probability of being S0 versus E (P_S0) (3)
--------------------------------------------------------------------------------
Note (1): proxy to clustered galaxies or projected pairs. Need for visual
inspection to select true on-going merger.
Note (2): TType values range = [-3.3, 8.0]
* ETG: TType ≤ 0
* LTG: Type > 0
Note (3): PS0 separates between Ell and S0. Only meaningful for T-Type≤0.
--------------------------------------------------------------------------------
History:
From electronic version of the journal
(End) Patricia Vannier [CDS] 27-Apr-2021