J/MNRAS/484/3879 Galaxy-scale lenses in the Kilo-Degree Survey (Petrillo+, 2019)
LinKS: discovering galaxy-scale strong lenses in the Kilo-Degree Survey using
convolutional neural networks.
Petrillo C.E., Tortora C., Vernardos G., Koopmans L.V.E.,
Verdoes Kleijn G., Bilicki M., Napolitano N.R., Chatterjee S., Covone G.,
Dvornik A., Erben T., Getman F., Giblin B., Heymans C., de Jong J.T.A.,
Kuijken K., Schneider P., Shan H., Spiniello C., Wright A.H.
<Mon. Not. R. Astron. Soc., 484, 3879-3896 (2019)>
=2019MNRAS.484.3879P 2019MNRAS.484.3879P (SIMBAD/NED BibCode)
ADC_Keywords: Gravitational lensing ; Galaxies ; Positional data
Keywords: gravitational lensing: strong -
galaxies: elliptical and lenticular, cD
Abstract:
We present a new sample of galaxy-scale strong gravitational lens
candidates, selected from 904deg2 of Data Release 4 of the
Kilo-Degree Survey, i.e. the 'Lenses in the Kilo-Degree Survey'
(LinKS) sample. We apply two convolutional neural networks (ConvNets)
to ∼88000 colour-magnitude-selected luminous red galaxies yielding a
list of 3500 strong lens candidates. This list is further downselected
via human inspection. The resulting LinKS sample is composed of 1983
rank-ordered targets classified as 'potential lens candidates' by at
least one inspector. Of these, a high-grade subsample of 89 targets is
identified with potential strong lenses by all inspectors.
Additionally, we present a collection of another 200 strong lens
candidates discovered serendipitously from various previous ConvNet
runs. A straightforward application of our procedure to future Euclid
or Large Synoptic Survey Telescope data can select a sample of ∼3000
lens candidates with less than 10 per cent expected false positives
and requiring minimal human intervention.
Description:
The Kilo-Degree Survey (KiDS; de Jong et al. 2013ExA....35...25D 2013ExA....35...25D) is
an ESO public survey carried out with the OmegaCAM wide-field imager
(Kuijken 2011Msngr.146....8K 2011Msngr.146....8K) mounted on the VLT Survey Telescope
(VST; Capaccioli & Schipani 2011Msngr.146....2C 2011Msngr.146....2C) at the Paranal
Observatory in Chile. The telescope, camera, and survey have been
designed to obtain images with subarcsecond seeing and homogeneous
image quality both across the full field of view and throughout the
survey execution.
To find gravitational lens candidates in KiDS imaging data, we use the
ConvNets previously introduced by Petrillo et al.
(2019MNRAS.482..807P 2019MNRAS.482..807P). These networks are significantly improved
variants of the original ConvNet presented by Petrillo et al.
(2017MNRAS.472.1129P 2017MNRAS.472.1129P). ConvNets represent a state-of-the-art method of
pattern recognition. The networks learn how to classify a diverse set
of images during the so-called training phase, whereby labelled images
are provided to the ConvNet. Its weight parameters are changed to
minimize a pre-defined loss function, which expresses the difference
between the labels of the images and the output values p (one for each
image) of the ConvNet.
We present in table A1 the 'Lenses in the Kilo-Degree Survey (LinKS)
sample' as the full sample of 1983 gravitational lens candidates
retrieved with p>0.8 and a score from the visual inspection greater
than zero.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
tablea1.dat 41 1983 LinKS sample
--------------------------------------------------------------------------------
Byte-by-byte Description of file: tablea1.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 4 I4 --- ID Internal lens identifier
6- 7 I2 --- Score Score from the visual inspection (1)
9- 13 F5.3 --- P1 ? p-value of the 1-band ConvNet (2)
15- 19 F5.3 --- P3 ? p-value of the 3-bands ConvNet (2)
21- 29 F9.5 deg RAdeg Right ascension (J2000)
31- 39 F9.5 deg DEdeg Declination (J2000)
41 I1 --- f_ID Flag indicating whether the candidate is
already a confirmed lens or it has been
identified as a candidate in other surveys (3)
--------------------------------------------------------------------------------
Note (1): The candidates are ordered by decreasing visual-inspection score
Note (2): Building on our experience, we choose to deploy two different
ConvNets. One focuses on utilizing the best morphological information
by taking the best seeing, i.e. r band, images as input. The other
ConvNet exploits colour information in addition to morphological
information by taking three-band RGB images as input. The ConvNets
take these images and transform them into a single value, p, which
can vary between 0 and 1. This value represents, to some degree,
the probability that the input image is a lens.
Note (3): Flag as follows:
0 = newly discovered candidates
1 = confirmed lenses
2 = SuGOHI lens candidates (see https://arxiv.org/abs/1704.01585 and
https://arxiv.org/abs/1809.07341)
--------------------------------------------------------------------------------
History:
From electronic version of the journal
(End) Ana Fiallos [CDS] 05-Sep-2022