Home > Research > Publications & Outputs > Lessons learned from the two largest Galaxy mor...

Research

Links

Text available via DOI:

View graph of relations

Lessons learned from the two largest Galaxy morphological classification catalogues built by convolutional neural networks

Research output: Contribution to Journal/Magazine › Journal article › peer-review

Published

Standard

Lessons learned from the two largest Galaxy morphological classification catalogues built by convolutional neural networks. / Dark Energy Survey Collaboration.
In: Monthly Notices of the Royal Astronomical Society, Vol. 518, No. 2, 31.01.2023, p. 2794-2809.

Research output: Contribution to Journal/Magazine › Journal article › peer-review

Harvard

Dark Energy Survey Collaboration 2023, 'Lessons learned from the two largest Galaxy morphological classification catalogues built by convolutional neural networks', Monthly Notices of the Royal Astronomical Society, vol. 518, no. 2, pp. 2794-2809. https://doi.org/10.1093/mnras/stac3228

APA

Dark Energy Survey Collaboration (2023). Lessons learned from the two largest Galaxy morphological classification catalogues built by convolutional neural networks. Monthly Notices of the Royal Astronomical Society, 518(2), 2794-2809. https://doi.org/10.1093/mnras/stac3228

Vancouver

Dark Energy Survey Collaboration. Lessons learned from the two largest Galaxy morphological classification catalogues built by convolutional neural networks. Monthly Notices of the Royal Astronomical Society. 2023 Jan 31;518(2):2794-2809. Epub 2022 Nov 11. doi: 10.1093/mnras/stac3228

Author

Dark Energy Survey Collaboration. / Lessons learned from the two largest Galaxy morphological classification catalogues built by convolutional neural networks. In: Monthly Notices of the Royal Astronomical Society. 2023 ; Vol. 518, No. 2. pp. 2794-2809.

Bibtex

@article{56abe79b04444d3b8c59c4bd3dc35dd6,
title = "Lessons learned from the two largest Galaxy morphological classification catalogues built by convolutional neural networks",
abstract = "We compare the two largest galaxy morphology catalogues, which separate early- and late-type galaxies at intermediate redshift. The two catalogues were built by applying supervised deep learning (convolutional neural networks, CNNs) to the Dark Energy Survey data down to a magnitude limit of ~21 mag. The methodologies used for the construction of the catalogues include differences such as the cutout sizes, the labels used for training, and the input to the CNN - monochromatic images versus gri-band normalized images. In addition, one catalogue is trained using bright galaxies observed with DES (i < 18), while the other is trained with bright galaxies (r < 17.5) and 'emulated' galaxies up to r-band magnitude 22.5. Despite the different approaches, the agreement between the two catalogues is excellent up to i < 19, demonstrating that CNN predictions are reliable for samples at least one magnitude fainter than the training sample limit. It also shows that morphological classifications based on monochromatic images are comparable to those based on gri-band images, at least in the bright regime. At fainter magnitudes, i > 19, the overall agreement is good (~95 per cent), but is mostly driven by the large spiral fraction in the two catalogues. In contrast, the agreement within the elliptical population is not as good, especially at faint magnitudes. By studying the mismatched cases, we are able to identify lenticular galaxies (at least up to i < 19), which are difficult to distinguish using standard classification approaches. The synergy of both catalogues provides an unique opportunity to select a population of unusual galaxies....",
author = "{Dark Energy Survey Collaboration} and Cheng, {T. -Y.} and {Dom{\'i}nguez S{\'a}nchez}, H. and J. Vega-Ferrero and Conselice, {C. J.} and M. Siudek and A. Arag{\'o}n-Salamanca and M. Bernardi and R. Cooke and L. Ferreira and M. Huertas-Company and J. Krywult and A. Palmese and A. Pieres and {Plazas Malag{\'o}n}, {A. A.} and {Carnero Rosell}, A. and D. Gruen and D. Thomas and D. Bacon and D. Brooks and James, {D. J.} and Hollowood, {D. L.} and D. Friedel and E. Suchyta and E. Sanchez and F. Menanteau and F. Paz-Chinch{\'o}n and G. Gutierrez and G. Tarle and I. Sevilla-Noarbe and I. Ferrero and J. Annis and J. Frieman and J. Garc{\'i}a-Bellido and J. Mena-Fern{\'a}ndez and K. Honscheid and K. Kuehn and {da Costa}, {L. N.} and M. Gatti and M. Raveri and Pereira, {M. E. S.} and M. Rodriguez-Monroy and M. Smith and {Carrasco Kind}, M. and M. Aguena and Swanson, {M. E. C.} and N. Weaverdyck and P. Doel and R. Miquel and Ogando, {R. L. C.} and Gruendl, {R. A.}",
year = "2023",
month = jan,
day = "31",
doi = "10.1093/mnras/stac3228",
language = "English",
volume = "518",
pages = "2794--2809",
journal = "Monthly Notices of the Royal Astronomical Society",
issn = "0035-8711",
publisher = "OXFORD UNIV PRESS",
number = "2",

}

RIS

TY - JOUR

T1 - Lessons learned from the two largest Galaxy morphological classification catalogues built by convolutional neural networks

AU - Dark Energy Survey Collaboration

AU - Cheng, T. -Y.

AU - Domínguez Sánchez, H.

AU - Vega-Ferrero, J.

AU - Conselice, C. J.

AU - Siudek, M.

AU - Aragón-Salamanca, A.

AU - Bernardi, M.

AU - Cooke, R.

AU - Ferreira, L.

AU - Huertas-Company, M.

AU - Krywult, J.

AU - Palmese, A.

AU - Pieres, A.

AU - Plazas Malagón, A. A.

AU - Carnero Rosell, A.

AU - Gruen, D.

AU - Thomas, D.

AU - Bacon, D.

AU - Brooks, D.

AU - James, D. J.

AU - Hollowood, D. L.

AU - Friedel, D.

AU - Suchyta, E.

AU - Sanchez, E.

AU - Menanteau, F.

AU - Paz-Chinchón, F.

AU - Gutierrez, G.

AU - Tarle, G.

AU - Sevilla-Noarbe, I.

AU - Ferrero, I.

AU - Annis, J.

AU - Frieman, J.

AU - García-Bellido, J.

AU - Mena-Fernández, J.

AU - Honscheid, K.

AU - Kuehn, K.

AU - da Costa, L. N.

AU - Gatti, M.

AU - Raveri, M.

AU - Pereira, M. E. S.

AU - Rodriguez-Monroy, M.

AU - Smith, M.

AU - Carrasco Kind, M.

AU - Aguena, M.

AU - Swanson, M. E. C.

AU - Weaverdyck, N.

AU - Doel, P.

AU - Miquel, R.

AU - Ogando, R. L. C.

AU - Gruendl, R. A.

PY - 2023/1/31

Y1 - 2023/1/31

N2 - We compare the two largest galaxy morphology catalogues, which separate early- and late-type galaxies at intermediate redshift. The two catalogues were built by applying supervised deep learning (convolutional neural networks, CNNs) to the Dark Energy Survey data down to a magnitude limit of ~21 mag. The methodologies used for the construction of the catalogues include differences such as the cutout sizes, the labels used for training, and the input to the CNN - monochromatic images versus gri-band normalized images. In addition, one catalogue is trained using bright galaxies observed with DES (i < 18), while the other is trained with bright galaxies (r < 17.5) and 'emulated' galaxies up to r-band magnitude 22.5. Despite the different approaches, the agreement between the two catalogues is excellent up to i < 19, demonstrating that CNN predictions are reliable for samples at least one magnitude fainter than the training sample limit. It also shows that morphological classifications based on monochromatic images are comparable to those based on gri-band images, at least in the bright regime. At fainter magnitudes, i > 19, the overall agreement is good (~95 per cent), but is mostly driven by the large spiral fraction in the two catalogues. In contrast, the agreement within the elliptical population is not as good, especially at faint magnitudes. By studying the mismatched cases, we are able to identify lenticular galaxies (at least up to i < 19), which are difficult to distinguish using standard classification approaches. The synergy of both catalogues provides an unique opportunity to select a population of unusual galaxies....

AB - We compare the two largest galaxy morphology catalogues, which separate early- and late-type galaxies at intermediate redshift. The two catalogues were built by applying supervised deep learning (convolutional neural networks, CNNs) to the Dark Energy Survey data down to a magnitude limit of ~21 mag. The methodologies used for the construction of the catalogues include differences such as the cutout sizes, the labels used for training, and the input to the CNN - monochromatic images versus gri-band normalized images. In addition, one catalogue is trained using bright galaxies observed with DES (i < 18), while the other is trained with bright galaxies (r < 17.5) and 'emulated' galaxies up to r-band magnitude 22.5. Despite the different approaches, the agreement between the two catalogues is excellent up to i < 19, demonstrating that CNN predictions are reliable for samples at least one magnitude fainter than the training sample limit. It also shows that morphological classifications based on monochromatic images are comparable to those based on gri-band images, at least in the bright regime. At fainter magnitudes, i > 19, the overall agreement is good (~95 per cent), but is mostly driven by the large spiral fraction in the two catalogues. In contrast, the agreement within the elliptical population is not as good, especially at faint magnitudes. By studying the mismatched cases, we are able to identify lenticular galaxies (at least up to i < 19), which are difficult to distinguish using standard classification approaches. The synergy of both catalogues provides an unique opportunity to select a population of unusual galaxies....

U2 - 10.1093/mnras/stac3228

DO - 10.1093/mnras/stac3228

M3 - Journal article

VL - 518

SP - 2794

EP - 2809

JO - Monthly Notices of the Royal Astronomical Society

JF - Monthly Notices of the Royal Astronomical Society

SN - 0035-8711

IS - 2

ER -