An image database: Automatic image analysis

Abstract

The LEDA extragalactic database is managing 35,841 images obtained by a direct ‘post-stamp’ digitization of the Palomar Sky Survey between δ = −33° and δ = 85°. These images are used to recognize automatically LEDA galaxies and to get their main parameters. So, a catalog of 28,00 galaxies is built. This process constitutes the first step of an new one which can lead to an automatic feeding of LEDA. The next step is the recognition of new galaxies. Some experiments have been done using Neural Network algorithm. These tests are briefly discussed.