J/A+A/668/A73        Search of strong lens systems in DES         (Rojas+, 2022)

Search of strong lens systems in the Dark Energy Survey using convolutional neural networks. Rojas K., Savary E., Clement B., Maus M., Courbin F., Lemon C., Chan J.H.H., Vernardos G., Joseph R., Canameras R., Galan A. <Astron. Astrophys. 668, A73 (2022)> =2022A&A...668A..73R 2022A&A...668A..73R (SIMBAD/NED BibCode)
ADC_Keywords: Surveys ; Gravitational lensing Keywords: gravitational lensing: strong - techniques: image processing - surveys - catalogs Abstract: We present our search for strong lens, galaxy-scale systems in the first data release of the Dark Energy Survey (DES), based on a color-selected parent sample of 18745029 luminous red galaxies (LRGs). We used a convolutional neural network (CNN) to grade this LRG sample with values between 0 (non-lens) and 1 (lens). Our training set of mock lenses is data-driven, that is, it uses lensed sources taken from HST-COSMOS images and lensing galaxies from DES images of our LRG sample. A total of 76582 cutouts were obtained with a score above 0.9, which were then visually inspected and classified into two catalogs. The first one contains 405 lens candidates, of which 90 present clear lensing features and counterparts, while the other 315 require more evidence, such as higher resolution imaging or spectra, to be conclusive. A total of 186 candidates are newly identified by our search, of which 16 are among the 90 most promising (best) candidates. The second catalog includes 539 ring galaxy candidates. This catalog will be a useful false positive sample for training future CNNs. For the 90 best lens candidates we carry out color-based deblending of the lens and source light without fitting any analytical profile to the data. This method is shown to be very efficient in the deblending, even for very compact objects and for objects with a complex morphology. Finally, from the 90 best lens candidates, we selected 52 systems with one single deflector to test an automated modeling pipeline that has the capacity to successfully model 79% of the sample within an acceptable computing runtime. Description: We used DES-DR1 to search for galaxy-scale strong lensing systems using a CNN that carries out a binary classification of optical images in the g, r; and i bands. In doing so, we targeted massive galaxies, that is, LRGs, which were selected using a wide color-magnitude cut accounting for realistic color contamination by the putative background star-forming blue galaxies. From our visual inspection, we created two main catalogs: a lens candidates catalog and a ring galaxy candidates catalog, the latter being our main source of contaminants. The first catalog contains a total of 405 lens candidate systems: 90 SL and 315 ML. Out of these, 186 were totally new systems and 219 were identified (but not necessarily confirmed) in previous searches. We deblended the lens and source light for our 90 SL systems using the MuSCADeT software, which does not involve any profile fitting, but uses the color contrast between the lens and source together with sparse regularization. This was successful in deblending most of the cases, where there were clear differences in the colors of the lens and source. The second catalog contains 539 ring galaxy candidates. We expect to use this ring catalog in the future to improve the training of machine learning algorithms in the recognition between lenses and ring galaxies. Still, 539 objects is not a sufficient sample for training CNNs and, thus, furthers work with, for instance, generative adversarial networks, is likely to be needed. Finally, we selected, from the SL category, the 52 systems that apparently had one well-defined galaxy as a deflector to test an automated modeling pipeline. File Summary: -------------------------------------------------------------------------------- FileName Lrecl Records Explanations -------------------------------------------------------------------------------- ReadMe 80 . This file table1.dat 101 90 Strong lens (SL) candidates table2.dat 96 315 Maybe lens (ML) candidates table3.dat 71 539 Ring galaxies candidates table4.dat 98 61 Strong lensing model refs.dat 65 30 References -------------------------------------------------------------------------------- See also: II/357 : The Dark Energy Survey (DES): Data Release 1 (Abbott+, 2018) II/371 : The Dark Energy Survey (DES): Data Release 2 (Abbott+, 2021) Byte-by-byte Description of file: table1.dat table2.dat -------------------------------------------------------------------------------- Bytes Format Units Label Explanations -------------------------------------------------------------------------------- 1- 18 A18 --- Cand Candidate name (DES JHHMMSS+DDMMSS) 20- 39 F20.16 deg RAdeg Right ascension (J2000) 41- 61 F21.17 deg DEdeg Declination (J2000) 63- 66 F4.2 --- SCnn [0.8/1.0] Cnn score (G1) 68- 71 F4.2 --- VISL [0/1] Visual inspection score (G1) 73-101 A29 --- Ref References (SL or ML) in refs.dat file -------------------------------------------------------------------------------- Byte-by-byte Description of file: table3.dat -------------------------------------------------------------------------------- Bytes Format Units Label Explanations -------------------------------------------------------------------------------- 1- 18 A18 --- Cand Candidate name (DES JHHMMSS+DDMMSS) 20- 39 F20.16 deg RAdeg Right ascension (J2000) 41- 61 F21.17 deg DEdeg Declination (J2000) 63- 66 F4.2 --- SCnn [0.9/1.0] Cnn score (G1) 68- 71 F4.2 --- VISR [0.28/1.0] Visual inspection score (G1) -------------------------------------------------------------------------------- Byte-by-byte Description of file: table4.dat -------------------------------------------------------------------------------- Bytes Format Units Label Explanations -------------------------------------------------------------------------------- 1- 18 A18 --- Cand Candidate name (DES JHHMMSS+DDMMSS) 19 A1 --- n_Cand [*] * for model result after manually redoing the mask 21- 24 F4.2 --- rchi2 Reduced chi2 of the model 26- 30 F5.3 arcsec RE Einstein radius 31 A1 --- --- [+] 32- 36 F5.3 arcsec E_RE Error on RE (upper value) 37 A1 --- --- [-] 38- 42 F5.3 arcsec e_RE Error on RE (lower value) 44- 48 F5.3 --- qm Axis ratio 49 A1 --- --- [+] 50- 54 F5.3 --- E_qm Error on qm ratio (upper value) 55 A1 --- --- [-] 56- 60 F5.3 --- e_qm Error on qm ratio (lower value) 62- 64 I3 deg PAm Position angle of the lens mass SIE model 65 A1 --- --- [+] 66- 67 I2 deg E_PAm Error on PAm (upper value) 68 A1 --- --- [-] 69- 70 I2 deg e_PAm Error on PAm (lower value) 72- 76 F5.3 --- gammaext strength of external shear 77 A1 --- --- [+] 78- 82 F5.3 --- E_gammaext Error on gammaext (upper value) 83 A1 --- --- [-] 84- 88 F5.3 --- e_gammaext Error on gammaext (lower value) 90- 92 I3 deg phiext angle of external shear 93 A1 --- --- [+] 94- 95 I2 deg E_phiext Error on phiext (upper value) 96 A1 --- --- [-] 97- 98 I2 deg e_phiext Error on phiext (lower value) -------------------------------------------------------------------------------- Byte-by-byte Description of file: refs.dat -------------------------------------------------------------------------------- Bytes Format Units Label Explanations -------------------------------------------------------------------------------- 1- 3 A3 --- Ref Reference code 5- 23 A19 --- BibCode BibCode 25- 42 A18 --- Aut Author's name 45- 65 A21 --- Com Comments -------------------------------------------------------------------------------- Global notes: Note (G1): images classified as lenses obtain ∼1 while non-lenses obtain ∼0. -------------------------------------------------------------------------------- History: From Karina Rojas, karina.rojasolate(at)gmail.com Acknowledgements: This work is supported by the Swiss National Science Foundation (SNSF) and by the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation program (COSMICLENS: grant agreement No 787886). RC thanks the Max Planck Society for support through the Max Planck Research Group of S.∼H.∼Suyu. This project has received funding from the European Research Council (ERC) under the European Unions Horizon 2020 research and innovation programme (LENSNOVA: grant agreement No 771776). GV has received funding from the European Union's Horizon 2020 research and innovation programme under the Marie Sklodovska-Curie grant agreement No 897124. This research has made use of the VizieR catalogue access tool, CDS, Strasbourg, France (DOI : 10.26093/cds/vizier). The original description of the VizieR service was published in Ochsenbein et al., 2000A&AS..143...23O 2000A&AS..143...23O . This research has made use of the SIMBAD database, operated at CDS, Strasbourg, France (Wenger et al., 2000A&AS..143....9W 2000A&AS..143....9W). This material is based upon work supported by the National Science Foundation under Cooperative Agreement 1258333 managed by the Association of Universities for Research in Astronomy (AURA), and the Department of Energy under Contract No. DE-AC02-76SF00515 with the SLAC National Accelerator Laboratory. Additional funding for Rubin Observatory comes from private donations, grants to universities, and in-kind support from LSSTC Institutional Members. This work rely on the following packages: numpy (van der Walt et al., 2011, Computing in Science Engineering, 13, 22), scipy (Virtanen et al., 2020, Nature Methods, 17, 261), matplotlib (Hunter, 2007, Computing in Science & Engineering, 9, 90), astropy (Prince-Whelam et al., 2018AJ....156..123A 2018AJ....156..123A) and aGithub scarlet_extensions (https://github.com/herjy/scarlet_extensions/releases/tag/paper) and csv. A notebook detailing the procedure of deblending can be found in the following repository: aGithub Lens-Deblend (https://github.com/herjy/Lens_deblend/releases/tag/prototype).
(End) Patricia Vannier [CDS] 10-Oct-2022
The document above follows the rules of the Standard Description for Astronomical Catalogues; from this documentation it is possible to generate f77 program to load files into arrays or line by line