A catalog of visual-like morphologies in the 5 candels fields using deep learning



Downloads per month over past year

Huertas Company, M. and Gravet, R. and Cabrera Vives, G. and Pérez González, Pablo Guillermo and Kartaltepe, J. S. and Barro, G. and Bernardi, M. and Mei, S. and Shankar, F. and Dimauro, P. and Bell, E. F. and Kocevski, D. and Koo, D. C. and Faber, S. M. and Mcintosh, D. H. (2015) A catalog of visual-like morphologies in the 5 candels fields using deep learning. Astrophysical journal supplement series, 221 (1). ISSN 0067-0049

[thumbnail of perezgonzalez157libre.pdf]

Official URL: http://dx.doi.org/10.1088/0067-0049/221/1/8


We present a catalog of visual-like H-band morphologies of ~50.000 galaxies (H_f160w < 24.5) in the 5 CANDELS fields (GOODS-N, GOODS-S, UDS, EGS, and COSMOS). Morphologies are estimated using Convolutional Neural Networks (ConvNets). The median redshift of the sample is < z> 1.25. The algorithm is trained on GOODS-S, for which visual classifications are publicly available, and then applied to the other 4 fields. Following the CANDELS main morphology classification scheme, our model retrieves for each galaxy the probabilities of having a spheroid or a disk, presenting an irregularity, being compact or a point source, and being unclassifiable. ConvNets are able to predict the fractions of votes given to a galaxy image with zero bias and ~10% scatter. The fraction of mis-classifications is less than 1%. Our classification scheme represents a major improvement with respect to Concentration-Asymmetry-Smoothness-based methods, which hit a 20%–30% contamination limit at high z.

Item Type:Article
Additional Information:

© 2015. The American Astronomical Society. All rights reserved. We thank the two anonymous referees for contributing to significantly improve this work. M.H.C acknowledges D. Gratadour for kindly giving us access to the GPU cluster at LESIA. G.C.V gratefully acknowledges financial support from CONICYT-Chile through its doctoral scholarship and grant DPI20140090. S.M. acknowledges financial support from the Institut Universitaire de France (IUF), of which she is senior member. G.B., D.C.K., and S.M.F. acknowledge support from NSF grant AST-08-08133 and NASA grant HST-GO-12060.10A.

Uncontrolled Keywords:High-redshift galaxies; Support vector machines; Seeing limited images; Digital sky survey; Neural-networks; Classification; Photometry; Evolution; Sequence
Subjects:Sciences > Physics > Astrophysics
Sciences > Physics > Astronomy
ID Code:35064
Deposited On:19 Jan 2016 14:11
Last Modified:10 Dec 2018 15:05

Origin of downloads

Repository Staff Only: item control page