Quantitative Morphology of Moderate Redshift Peculiar Galaxies

Abstract

We report the results of training an Artificial Neural Network (ANN) classifier to distinguish between peculiar and normal galaxies on Hubble Space Telescope Wide Field Planetary Camera 2 images. 978 I band (filter F814W) images from 9 Groth-Westphal strip fields were classified by eye into five broad types (E/SO; early S; late S and two peculiar types). We examine the light concentration and asymmetry parameters (the C-A set) suggested by Abraham et al. (1995), as well as design a set of four morphological parameters (the 4P set) to describe each galaxy. Our parameters are:

• 1 Overall Texture (or “blobbiness”): The degree of departure from a smooth, radially decreasing light distribution. Picks up bright localised structures.

• 2 The distortion of isophotes: The distance between the geometrical centers of different isophotes. Indicates overall distortions.

• 3 The filling-factor of isophotes: The degree to which regions enclosing isophotes are filled by pixels of that isophotal level. Indicates the existence of structures.

• 4 Skeleton ratios of detected structures: the ratio between the size of the skeleton and that of the corresponding region. Measures elongation of structures.