astrophysics

Automated Computer Vision Cluster Identification in the Fireworks Galaxy

arXiv:2607.26330

summary

The paper describes a computer‑vision method that uses Hubble Space Telescope UV images to automatically detect and measure young star cluster candidates in the Fireworks Galaxy (NGC 6946), recovering about 60% of synthetic clusters with a modest false‑positive rate.

Abstract

We present the integrated photometry, radii, and spatial distribution of young ( 25 Myr) star cluster candidates in NGC 6946. NGC 6946, also known as the Fireworks Galaxy, is a highly star-forming galaxy with numerous young massive clusters. We have developed a modified computer vision algorithm using photometry from images taken with Hubble Space Telescope (HST) Wide Field Camera 3 Ultraviolet channel (WFC3/UVIS) F275W and F336W filters to identify and outline candidate clusters. We describe our technique in detail, including extensive testing with artificial clusters, where the algorithm recovers 60.7% of synthetic clusters and has a conservative false positive rate of 27.3% down to luminosities of M. We identify 6410 cluster candidates down to much fainter magnitudes (M) via the aforementioned algorithm which are more difficult to verify, but are still of interest as the luminosity function of these candidates is consistent with a standard power law with a slope of 2.

15 pages, 9 figures; accepted for publication in The Astrophysical Journal

Topics & keywords

#star clusters#computer vision#galactic astronomy#Hubble Space Telescope#photometric analysisWFC3/UVISF275WF336Wsynthetic cluster recoveryluminosity functionpower‑law slope
Automated Computer Vision Cluster Identification in the Fireworks Galaxy · wovepaper