O'TRAIN: A robust and flexible 'real or bogus' classifier for the study of the optical transient sky

0568212 - FZÚ 2023 RIV FR eng J - Journal Article
Makhlouf, K. - Turpin, D. - Corre, D. - Karpov, Sergey - Kann, D.A. - Klotz, A.
O'TRAIN: A robust and flexible 'real or bogus' classifier for the study of the optical transient sky.
Astronomy & Astrophysics. Roč. 664, Aug (2022), č. článku A81. ISSN 0004-6361. E-ISSN 1432-0746
R&D Projects: GA MŠMT EF15_003/0000437; GA MŠMT(CZ) LM2018105; GA MŠMT LM2015046; GA MŠMT LTT17006; GA MŠMT EF16_013/0001403; GA MŠMT(CZ) EF18_046/0016007
Grant - others:OP VVV - CoGraDS(XE) CZ.02.1.01/0.0/0.0/15_003/0000437; OP VVV - CTAO-CZ(XE) CZ.02.1.01/0.0/0.0/16_013/0001403; OP VVV - CTA-CZ II(XE) CZ.02.1.01/0.0/0.0/18_046/0016007
Institutional support: RVO:68378271
Keywords : methods:numerical * techniques: image processing * Astrophysics - Instrumentation and Methods for Astrophysics
OECD category: Astronomy (including astrophysics,space science)
Impact factor: 6.5, year: 2022
Method of publishing: Open access

Deep machine learning algorithms have now proven their efficiency in recognising patterns in images. These methods are now used in astrophysics to perform different classification tasks such as identifying bogus from real transient point-like sources. We explore this method to provide a robust and flexible algorithm that could be included in any kind of transient detection pipeline. We built a convolutional neural network (CNN) algorithm in order to perform a `real or bogus' classification task on transient candidate cutouts (subtraction residuals) provided by different kinds of optical telescopes. The training involved human-supervised labelling of the cutouts, which are split into two balanced data sets with `true' and `false' point-like source candidates. We tested our CNN model on the candidates produced by two different transient detection pipelines. In addition, we made use of several diagnostic tools to evaluate the classification performance of our CNN models.
Permanent Link: https://hdl.handle.net/11104/0339546