Capivara: A Spectral-based Segmentation Method for IFU Data Cubes

We present capivara, a fast and scalable multi-decomposition package designed to study astrophysical properties within distinct structural components of galaxies. Our spectro-decomposition code for analyzing integral field unit (IFU) data enables a more holistic approach, moving beyond conventional...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2024-10
Main Authors: de Souza, Rafael S, Dahmer-Hahn, Luis G, Shen, Shiyin, Chies-Santos, Ana L, Chen, Mi, Rahna, P T, Ye, Renhao, Tahmasebzade, Behzad
Format: Article
Language:English
Subjects:
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We present capivara, a fast and scalable multi-decomposition package designed to study astrophysical properties within distinct structural components of galaxies. Our spectro-decomposition code for analyzing integral field unit (IFU) data enables a more holistic approach, moving beyond conventional radial gradients and the bulge-plus-disk dichotomy. It facilitates comprehensive comparisons of integrated stellar ages and metallicities across various galactic structures. Our classification method naturally identifies outliers and organizes the different pixels based on their dominant spectral features. The algorithm leverages the scalability and GPU acceleration of Torch, outputting both a one-dimensional spectrum and a full data cube for each galaxy component, without relying on Voronoi binning. We demonstrate the capabilities of our approach using a sample of galaxies from the MaNGA survey, processing the resulting data cubes with the starlight spectral fitting code to derive both stellar population and ionized gas properties of the galaxy components. Our method effectively groups regions with similar spectral properties in both the continuum and emission lines. By aggregating the spectra of these regions, we enhance the signal-to-noise ratio of our analysis while significantly speeding up computations by reducing the number of spectra processed simultaneously. capivara will be freely available on GitHub.
ISSN:2331-8422