EUVIMEDEUVIMED entdecken
Crossref

Using Image Processing and Deep Learning to Detect Dwarf Galaxies in NGC 5128

Cameron Leahy

Peer Beyond Graduate Research Conference · 19.05.2026 · 2 (1)

DOI: 10.55016/bhst3h52 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Peer Beyond Graduate Research Conference, online erschienen am 19.05.2026. Verfasst von Cameron Leahy.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

NGC 5128, also known as Centaurus A, is the closest giant galaxy beyond the Milky Way and Andromeda (Harris et al., 2010). It is known to host a rich population of dwarf galaxy satellites, with many dwarfs waiting to be discovered (Taylor et al. 2018). In my research, I design an algorithm that uses novel image processing techniques and a custom-built convolutional neural network (CNN) to perform automatic detection of dwarf galaxy candidates in optical Dark Energy Camera imaging (DECam; Flaugher et al. 2015). While image processing routines and CNNs have been used individually in faint dwarf galaxy detection before (Bennet et al. 2017, Tanoglidis et al. 2021), the fusion of both techniques into a data-reduction pipeline for DECam imaging represents a novel contribution. The dwarf galaxy candidates uncovered by the algorithm contribute to the ongoing analysis of the NGC 5128 dwarf galaxy system and serve as a valuable testing ground for cosmological theory.

Zitieren

APA

Leahy, C. (2026). Using Image Processing and Deep Learning to Detect Dwarf Galaxies in NGC 5128. Peer Beyond Graduate Research Conference, 2(1). https://doi.org/10.55016/bhst3h52

Vancouver

Leahy C. Using Image Processing and Deep Learning to Detect Dwarf Galaxies in NGC 5128. Peer Beyond Graduate Research Conference. 2026;2(1). doi:10.55016/bhst3h52

BibTeX
@article{Leahy2026,
  author = {Leahy, Cameron},
  title = {Using Image Processing and Deep Learning to Detect Dwarf Galaxies in NGC 5128},
  journal = {Peer Beyond Graduate Research Conference},
  year = {2026},
  volume = {2},
  number = {1},
  doi = {10.55016/bhst3h52}
}
Kostenlos starten →