Support Vector Machines and CASGM20 Parameters Applied to Morphological Classification of Reconstructed 2D Images of Extended Objects Within the ESA-Gaia Mission

被引：0

作者：

Krone-Martins, Alberto ^{[1
,2
]}

Ducourant, Christine ^{[1
]}

Teixeira, Ramachrisna ^{[2
]}

机构：

[1] Lab Astrophys Bordeaux, 2 Rue Observ,BP 89, F-33271 Floirac, France

[2] Univ Sao Paulo, Inst Astron Geofis & Ciencias Atmosfer, BR-05508090 Sao Paulo, SP, Brazil

来源：

CLASSIFICATION AND DISCOVERY IN LARGE ASTRONOMICAL SURVEYS | 2008年 / 1082卷

基金：

巴西圣保罗研究基金会;

关键词：

Classification and classification systems; Morphology and overall structure; Mathematical procedures and computer techniques; Astrometry;

D O I：

暂无

中图分类号：

P1 [天文学];

学科分类号：

0704 ;

摘要：

In this work we present some parameters that ate being studied to perform a purely morphological analysis of reconstructed images of extended objects, particularly galaxies, in the context of the ESA-Gaia mission. Those parameters, known as Concentration, Asymmetry, Clumpiness, Gini's coefficient and the Momentum of the brightest 20% of the galaxy, form a set that is becoming commonly used when a limited number of pixels is available to analyse, such as will be the case for Gaia reconstructed images. We comment about small modifications on those parameters that are planned to be performed. We also report tests with a preliminar version of the code that is being written to analyse Crtia images on a sample based on the Frei catalog of galaxies. Finally, we comment on the possibility of using Support Vector Machines to perform the morphological classification based on those measured parameters, and conclude that a very good level of segregation can be obtained for a two-class discrimination.

引用

页码：151 / +

页数：2