Searching for Edge-on galaxies in the DESI Legacy surveys using ANN

Examples

The DESI Legacy Surveys covers 20,000 square degrees of extragalactic sky combining MzLS, DECaLS and BASS surveys and public DECam data in g, r, i, z bands. The goal of the project is to create a robust sample of edge-on galaxies covering the entire sky outside the Zone of Avoidance in the Milky Way. The edge-on galaxies are selected using an artificial neural network trained on the sample of the edge-on galaxies from the second version of the Edge-on Galaxies In the Pan-STARRS1 survey (EGIPS) and the Revised Flat Galaxy Catalogue (RFGC).

Team

Alexandra Antipova (SAO RAS, Pulkovo Observatory)
Dmitry Bizyaev (APO, SAI MSU)
Svyatoslav Borisov
Ilia Chugunov (Pulkovo Observatory)
Dmitry Makarov (SAO RAS, Pulkovo Observatory)
Alexander Marchuk (Pulkovo Observatory, SPbU)
Alexandra Nazarova (SAO RAS, Pulkovo Observatory)
Vladimir Reshetnikov (SPbU, Pulkovo Observatory)
Evgeny Rubtsov (SAI MSU)
Sergey Savchenko (SPbU, Pulkovo Observatory)
Daniil Smirnov
Anastasia Sypkova (SPbU, Pulkovo Observatory)
Iliya Tikhonenko

Statistics on the Edge-on / Non-Edge-on separation

Based on personal perceptions of edge-on galaxies, people can be divided into groups: 'bad cops' - who judge the galaxies strictly, 'good cops' - who use soft criteria, and a gradual transition between these extreme cases. The 'bad cops' rejected about 40% of objects, while the 'good cops' excluded only 10%.

Statistics of the number of votes given for a galaxy being edge-on, depending on the level of agreement (percentage ranges).