A Caltech Library Service

Star cluster classification in the PHANGS-HST survey: Comparison between human and machine learning approaches

Whitmore, Bradley C. and Lee, Janice C. and Chandar, Rupali and Thilker, David A. and Hannon, Stephen and Wei, Wei and Huerta, E. A. and Bigiel, Frank and Boquien, Médéric and Chevance, Mélanie and Dale, Daniel A. and Deger, Sinan and Grasha, Kathryn and Klessen, Ralf S. and Kruijssen, J. M. Diederik and Larson, Kirsten L. and Mok, Angus and Rosolowsky, Erik and Schinnerer, Eva and Schruba, Andreas and Ubeda, Leonardo and Van Dyk, Schuyler D. and Watkins, Elizabeth and Williams, Thomas (2021) Star cluster classification in the PHANGS-HST survey: Comparison between human and machine learning approaches. Monthly Notices of the Royal Astronomical Society, 506 (4). pp. 5294-5317. ISSN 0035-8711. doi:10.1093/mnras/stab2087.

[img] PDF - Published Version
See Usage Policy.

[img] PDF - Accepted Version
Creative Commons Attribution.


Use this Persistent URL to link to this item:


When completed, the PHANGS–HST project will provide a census of roughly 50 000 compact star clusters and associations, as well as human morphological classifications for roughly 20 000 of those objects. These large numbers motivated the development of a more objective and repeatable method to help perform source classifications. In this paper, we consider the results for five PHANGS–HST galaxies (NGC 628, NGC 1433, NGC 1566, NGC 3351, NGC 3627) using classifications from two convolutional neural network architectures (RESNET and VGG) trained using deep transfer learning techniques. The results are compared to classifications performed by humans. The primary result is that the neural network classifications are comparable in quality to the human classifications with typical agreement around 70 to 80 per cent for Class 1 clusters (symmetric, centrally concentrated) and 40 to 70 per cent for Class 2 clusters (asymmetric, centrally concentrated). If Class 1 and 2 are considered together the agreement is 82 ± 3 per cent. Dependencies on magnitudes, crowding, and background surface brightness are examined. A detailed description of the criteria and methodology used for the human classifications is included along with an examination of systematic differences between PHANGS–HST and LEGUS. The distribution of data points in a colour–colour diagram is used as a ‘figure of merit’ to further test the relative performances of the different methods. The effects on science results (e.g. determinations of mass and age functions) of using different cluster classification methods are examined and found to be minimal.

Item Type:Article
Related URLs:
URLURL TypeDescription Paper ItemMikulski Archive for Space Telescopes ItemPHANGS-HST
Whitmore, Bradley C.0000-0002-3784-7032
Lee, Janice C.0000-0002-2278-9407
Chandar, Rupali0000-0003-0085-4623
Thilker, David A.0000-0002-8528-7340
Hannon, Stephen0000-0001-9628-8958
Wei, Wei0000-0002-1018-7708
Huerta, E. A.0000-0002-9682-3604
Bigiel, Frank0000-0003-0166-9745
Boquien, Médéric0000-0003-0946-6176
Chevance, Mélanie0000-0002-5635-5180
Dale, Daniel A.0000-0002-5782-9093
Deger, Sinan0000-0003-1943-723X
Grasha, Kathryn0000-0002-3247-5321
Klessen, Ralf S.0000-0002-0560-3172
Kruijssen, J. M. Diederik0000-0002-8804-0212
Larson, Kirsten L.0000-0003-3917-6460
Mok, Angus0000-0001-7413-7534
Rosolowsky, Erik0000-0002-5204-2259
Schinnerer, Eva0000-0002-3933-7677
Van Dyk, Schuyler D.0000-0001-9038-9950
Watkins, Elizabeth0000-0002-7365-5791
Williams, Thomas0000-0002-0012-2142
Additional Information:© 2021 The Author(s). Published by Oxford University Press on behalf of Royal Astronomical Society. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model ( Accepted 2021 July 16. Received 2021 July 16; in original form 2021 March 16. Published: 21 July 2021. We thank the referee for a number of insightful comments that we feel have greatly improved the paper. This study is based on observations made with the NASA/ESA Hubble Space Telescope, obtained from the data archive at the Space Telescope Science Institute. STScI is operated by the Association of Universities for Research in Astronomy, Inc. under NASA contract NAS5-26555. Support for Program number 15654 was provided through a grant from the STScI under NASA contract NAS5-26555. JMDK and MC gratefully acknowledge funding from the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) through an Emmy Noether Research Group (grant number KR4801/1-1) and the DFG Sachbeihilfe (grant number KR4801/2-1). JMDK gratefully acknowledges funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme via the ERC Starting Grant MUSTANG (grant agreement number 714907). TGW acknowledges funding from the European Research Council (ERC) under the European Unionś Horizon 2020 research and innovation programme (grant agreement No. 694343). EAH and WW gratefully acknowledge National Science Foundation (NSF) awards OAC-1931561 and OAC-1934757. FB acknowledges funding from the European Research Council (ERC) under the European Unions Horizon 2020 research and innovation programme (grant agreement No. 726384/Empire). Data Availability: The data underlying this article are available at the Mikulski Archive for Space Telescopes at under proposal GO-15654. High level science products associated with HST GO-15654 are provided at
Group:Infrared Processing and Analysis Center (IPAC), TAPIR
Funding AgencyGrant Number
Deutsche Forschungsgemeinschaft (DFG)KR4801/1-1
Deutsche Forschungsgemeinschaft (DFG)KR4801/2-1
European Research Council (ERC)714907
European Research Council (ERC)694343
European Research Council (ERC)726384
Subject Keywords:catalogues – galaxies: star clusters: general
Issue or Number:4
Record Number:CaltechAUTHORS:20211110-172515036
Persistent URL:
Official Citation:Bradley C Whitmore, Janice C Lee, Rupali Chandar, David A Thilker, Stephen Hannon, Wei Wei, E A Huerta, Frank Bigiel, Médéric Boquien, Mélanie Chevance, Daniel A Dale, Sinan Deger, Kathryn Grasha, Ralf S Klessen, J M Diederik Kruijssen, Kirsten L Larson, Angus Mok, Erik Rosolowsky, Eva Schinnerer, Andreas Schruba, Leonardo Ubeda, Schuyler D Van Dyk, Elizabeth Watkins, Thomas Williams, Star cluster classification in the PHANGS–HST survey: Comparison between human and machine learning approaches, Monthly Notices of the Royal Astronomical Society, Volume 506, Issue 4, October 2021, Pages 5294–5317,
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:111826
Deposited By: Tony Diaz
Deposited On:11 Nov 2021 18:46
Last Modified:11 Nov 2021 18:46

Repository Staff Only: item control page