AutoSourceID-Classifier - Star-galaxy classification using a convolutional neural network with spatial information

Open Access

Table 1

Performance metrics by S/N.

S/N	~ 7	~ 13	~ 26	~ 50	~ 96
AUPRC ASID (s)	0.952	0.990	0.998	1.000	1.000
AUPRC SE (s)	0.841	0.975	0.998	1.000	1.000
AUPRC Base (s)	0.813	0.893	0.950	0.984	0.997

AUPRC ASID (g)	0.614	0.786	0.899	0.937	0.947
AUPRC SE (g)	0.271	0.575	0.782	0.853	0.782
AUPRC Base (g)	0.187	0.107	0.050	0.016	0.003

Brier ASID	0.108	0.044	0.012	0.002	0.000
Brier SE	0.279	0.188	0.076	0.034	0.015
Brier Base	0.250	0.185	0.140	0.113	0.101

Notes. The AUPRC, due to its formulation, is calculated separately for both stars (s) and galaxies (g) as the positive class. The table shows that ASID-C consistently outperforms both SE and the baseline model across most S/N values and metrics.

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.