Search NASA⌕ Search

Engineering topics

Thilker, David A.

Publications and source records attributed to Thilker, David A..

Star cluster classification using deep transfer learning with PHANGS- HST

Currently available star cluster catalogues from the Hubble Space Telescope (HST) imaging of nearby galaxies heavily rely on visual inspection and classification of candidate clusters. Here, the time-consuming nature of this process has limited the production of reliable catalogues and thus also post-observation analysis. To address this problem, deep transfer learning has recently been used to create neural network models that accurately classify star cluster morphologies at production scale for nearby spiral galaxies (D ≲ 20 Mpc). Here, we use HST ultraviolet (UV)–optical imaging of over 20 000 sources in 23 galaxies from the Physics at High Angular resolution in Nearby GalaxieS (PHANGS) survey to train and evaluate two new sets of models: (i) distance-dependent models, based on cluster candidates binned by galaxy distance (9–12, 14–18, and 18–24 Mpc), and (ii) distance-independent models, based on the combined sample of candidates from all galaxies. We find that the overall accuracy of both sets of models is comparable to previous automated star cluster classification studies (~60–80 per cent) and shows improvement by a factor of 2 in classifying asymmetric and multipeaked clusters from PHANGS-HST. Somewhat surprisingly, while we observe a weak negative correlation between model accuracy and galactic distance, we find that training separate models for the three distance bins does not significantly improve classification accuracy. We also evaluate model accuracy as a function of cluster properties such as brightness, colour, and spectral energy distribution (SED)-fit age. Based on the success of these experiments, our models will provide classifications for the full set of PHANGS-HST candidate clusters (N ~ 200 000) for public release.

79 ASTRONOMY AND ASTROPHYSICS↗

Deep transfer learning for star cluster classification: I. application to the PHANGS– HST survey

ABSTRACT We present the results of a proof-of-concept experiment that demonstrates that deep learning can successfully be used for production-scale classification of compact star clusters detected in Hubble Space Telescope(HST) ultraviolet-optical imaging of nearby spiral galaxies ($D\lesssim 20\, \textrm{Mpc}$) in the Physics at High Angular Resolution in Nearby GalaxieS (PHANGS)–HST survey. Given the relatively small nature of existing, human-labelled star cluster samples, we transfer the knowledge of state-of-the-art neural network models for real-object recognition to classify star clusters candidates into four morphological classes. We perform a series of experiments to determine the dependence of classification performance on neural network architecture (ResNet18 and VGG19-BN), training data sets curated by either a single expert or three astronomers, and the size of the images used for training. We find that the overall classification accuracies are not significantly affected by these choices. The networks are used to classify star cluster candidates in the PHANGS–HST galaxy NGC 1559, which was not included in the training samples. The resulting prediction accuracies are 70 per cent, 40 per cent, 40–50 per cent, and 50–70 per cent for class 1, 2, 3 star clusters, and class 4 non-clusters, respectively. This performance is competitive with consistency achieved in previously published human and automated quantitative classification of star cluster candidate samples (70–80 per cent, 40–50 per cent, 40–50 per cent, and 60–70 per cent). The methods introduced herein lay the foundations to automate classification for star clusters at scale, and exhibit the need to prepare a standardized data set of human-labelled star cluster classifications, agreed upon by a full range of experts in the field, to further improve the performance of the networks introduced in this study.

Wei, Wei↗

THE STRUCTURE AND STELLAR CONTENT OF THE OUTER DISKS OF GALAXIES: A NEW VIEW FROM THE Pan-STARRS1 MEDIUM DEEP SURVEY

We present the results of an analysis of Pan-STARRS1 Medium Deep Survey multi-band (grizy) images of a sample of 698 low-redshift disk galaxies that span broad ranges in stellar mass, star-formation rate, and bulge/disk ratio. We use population synthesis spectral energy distribution fitting techniques to explore the radial distribution of the light, color, surface mass density, mass/light ratio, and age of the stellar populations. We characterize the structure and stellar content of the galaxy disks out to radii of about twice Petrosian r {sub 90}, beyond which the halo light becomes significant. We measure normalized radial profiles for sub-samples of galaxies in three bins each of stellar mass and concentration. We also fit radial profiles to each galaxy. The majority of galaxies have down-bending radial surface brightness profiles in the bluer bands with a break radius at roughly r {sub 90}. However, they typically show single unbroken exponentials in the reddest bands and in the stellar surface mass density. We find that the mass/light ratio and stellar age radial profiles have a characteristic 'U' shape. There is a good correlation between the amplitude of the down-bend in the surface brightness profile and the rate of the increase in the M/L ratio in the outer disk. As we move from late- to early-type galaxies, the amplitude of the down-bend and the radial gradient in M/L both decrease. Our results imply a combination of stellar radial migration and suppression of recent star formation can account for the stellar populations of the outer disk.

79 ASTRONOMY AND ASTROPHYSICS↗

QUANTIFYING NON-STAR-FORMATION-ASSOCIATED 8 {mu}m DUST EMISSION IN NGC 628

Combining H{alpha} and IRAC images of the nearby spiral galaxy NGC 628, we find that between 30% and 43% of its 8 {mu}m dust emission is not related to recent star formation. Contributions from dust heated by young stars are separated by identifying H II regions in the H{alpha} map and using these areas as a mask to determine the 8 {mu}m dust emission that must be due to heating by older stars. Corrections are made for sub-detection-threshold H II regions, photons escaping from H II regions, and for young stars not directly associated with H II regions (i.e., 10-100 Myr old stars). A simple model confirms that this amount of 8 {mu}m emission can be expected given dust and PAH absorption cross sections, a realistic star formation history, and the observed optical extinction values. A Fourier power spectrum analysis indicates that the 8 {mu}m dust emission is more diffuse than the H{alpha} emission (and similar to observed H I), supporting our analysis that much of the 8 {mu}m-emitting dust is heated by older stars. The 8 {mu}m dust-to-H{alpha} emission ratio declines with galactocentric radius both within and outside of H II regions, probably due to a radial increase in disk transparency. In the course of this work, we have also found that intrinsic diffuse H{alpha} fractions may be lower than previously thought in galaxies, if the differential extinction between H II regions and diffuse regions is taken into account.

79 ASTRONOMY AND ASTROPHYSICS↗