Project EFIGI: Automatic classification of galaxies

From LRDE

Revision as of 10:54, 17 April 2015 by Bot (talk | contribs)

Abstract

We propose an automatic system to classify images of galaxies with varying resolution. Morphologically typing galaxies is a difficult task in particular for distant galaxies convolved by a point-spread function and suffering from a poor signal-to-noise ratio. In the context of the first phase of the project EFIGI (extraction of the idealized shapes of galaxies in imagery), we present the three steps of our software: cleaning, dimensionality reduction and supervised learning. We present preliminary results derived from a subset of 774 galaxies from the Principal Galaxies Catalog and compare them to human classifications made by astronomers. We use g-band images from the Sloan Digital Sky Survey. Finally, we discuss future improvements which we intend to implement before releasing our tool to the community.


Bibtex (lrde.bib)

@InProceedings{	  baillard.05.adass,
  author	= {Anthony Baillard and Emmanuel Bertin and Yannic Mellier
		  and Henry Joy {McCracken} and Thierry G\'eraud and Roser
		  Pell\'o and Jean-Fran\c{c}ois {LeBorgne} and Pascal Fouqu\'e},
  title		= {Project {EFIGI}: Automatic classification of galaxies},
  year		= 2005,
  booktitle	= {Astronomical Data Analysis Software and Systems XV},
  volume	= 351,
  pages		= {236--239},
  publisher	= {Astronomical Society of the Pacific},
  series	= {Conference},
  url		= {http://www.aspbooks.org/custom/publications/paper/index.phtml?paper_id=3398},
  editor	= {Carlos Gabriel and Christophe Arviset and Daniel Ponz and
		  Enrique Solano},
  isbn		= {1-58381-219-9},
  project	= {Image},
  abstract	= {We propose an automatic system to classify images of
		  galaxies with varying resolution. Morphologically typing
		  galaxies is a difficult task in particular for distant
		  galaxies convolved by a point-spread function and suffering
		  from a poor signal-to-noise ratio. In the context of the
		  first phase of the project EFIGI (extraction of the
		  idealized shapes of galaxies in imagery), we present the
		  three steps of our software: cleaning, dimensionality
		  reduction and supervised learning. We present preliminary
		  results derived from a subset of 774 galaxies from the
		  Principal Galaxies Catalog and compare them to human
		  classifications made by astronomers. We use g-band images
		  from the Sloan Digital Sky Survey. Finally, we discuss
		  future improvements which we intend to implement before
		  releasing our tool to the community.}
}