Star Cluster Classification in the PHANGS-HST Survey: Comparison between Human and Machine Learning Approaches

Abstract

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$\%$ for Class 1 clusters (symmetric, centrally concentrated) and 40 to 70$\%$ for Class 2 clusters (asymmetric, centrally concentrated). If Class 1 and 2 are considered together the agreement is 82 $\pm$ 3$\%$. 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.

28 pages, 25 figures. Accepted for publication in Monthly Notices of the Royal Astronomical Society. Version with full resolution figures found at https://sites.google.com/view/phangs/publications

References in corpus (19)

• A comprehensive set of simulations studying the influence of gas expulsion on star cluster evolution
• The distinction between star clusters and associations
• Star clusters near and far; tracing star formation across cosmic time
• ALMA Observations of the Antennae Galaxies: I. A New Window on a Prototypical Merger
• The Spatial Relation between Young Star Clusters and Molecular Clouds in M 51 with LEGUS
• OB associations and their origins
• PHAT Stellar Cluster Survey I. Year 1 Catalog and Integrated Photometry
• PHAT Stellar Cluster Survey. II. Andromeda Project Cluster Catalog
• Not all stars form in clusters - measuring the kinematics of OB associations with $Gaia$
• H$\alpha$ Morphologies of Star Clusters: A LEGUS study of HII region evolution timescales and stochasticity in low mass clusters
• The Young Star Cluster population of M51 with LEGUS: I. A comprehensive study of cluster formation and evolution
• Star Cluster Catalogs for the LEGUS Dwarf Galaxies
• PHANGS-HST: Star Cluster Spectral Energy Distribution Fitting with CIGALE
• LEGUS and Halpha-LEGUS Observations of Star Clusters in NGC 4449: Improved Ages and the Fraction of Light in Clusters as a Function of Age
• Star Cluster Mass and Age Distributions of Two Fields in M83 Based on HST/WFC3 Observations
• Biases on initial mass function determinations. III. Cluster masses derived from unresolved photometry
• StarcNet: Machine Learning for Star Cluster Identification
• The Three Young Nuclear Super Star Clusters in NGC 5253
• The effect of spatial resolution on optical and near-IR studies of stellar clusters: Implications for the origin of the red excess