Poster Presentation
Calibrating Non-Parametric Morphology with Citizen-Science Labels: New Gini–M20 and CAS Boundaries for HSC Galaxies
Presenter: YUEYI Kang (Ritsumeikan University)
Galaxy morphology provides essential clues for understanding galaxy formation and evolution, as well as the growth of large-scale structures in the Universe. With the rapid advancement of observational techniques, contemporary studies of galaxy morphological classification increasingly employ a set of scale-invariant parameters to quantify galaxy structure. Among these, concentration (C), asymmetry (A), smoothness (S), Gini coefficient (G), and second-order moment of the brightest 20% of the flux (M20) have been widely used for morphological classification across multiple wavelengths.
In this work, we used i-band imaging data from the HSC-SSP, combined with visual classifications from the citizen-science project GALAXY CRUISE (season 1: Tanaka et al. 2023). Using 19,276 galaxies from the HSC s23a wide-layer data, we refitted the decision boundary separating spiral and elliptical galaxies on the Gini–M20 plane using the Python-based morphological analysis tool statmorph (Rodriguez-Gomez et al. 2019).
We first recalibrated the boundary between elliptical and spiral galaxies for non-merging systems using the merger/non-merger classification proposed by Lotz (2008) as a reference. We then derived a new early-/late-type classification boundary applicable to the entire sample, including galaxies identified as mergers. In addition, using the CAS system, we identified a set of threshold conditions suitable for selecting the spiral galaxies.
Our results show that the new G–M20 boundary achieves a 75% classification accuracy for galaxies with P(Spiral) > 0.8, whereas the newly defined CAS-based thresholds yield an improved accuracy of 85% for the same sample.
asiaa.sinica.edu.tw