TY - JOUR
T1 - Tails
T2 - Chasing Comets with the Zwicky Transient Facility and Deep Learning
AU - Duev, Dmitry A.
AU - Bolin, Bryce T.
AU - Graham, Matthew J.
AU - Kelley, Michael S.P.
AU - Mahabal, Ashish
AU - Bellm, Eric C.
AU - Coughlin, Michael W.
AU - Dekany, Richard
AU - Helou, George
AU - Kulkarni, Shrinivas R.
AU - Masci, Frank J.
AU - Prince, Thomas A.
AU - Riddle, Reed
AU - Soumagnac, Maayane T.
AU - Van Der Walt, Stéfan J.
N1 - Publisher Copyright:
© 2021. The American Astronomical Society. All rights reserved.
PY - 2021/5
Y1 - 2021/5
N2 - We present Tails, an open-source deep-learning framework for the identification and localization of comets in the image data of the Zwicky Transient Facility (ZTF), a robotic optical time-domain survey currently in operation at the Palomar Observatory in California, USA. Tails employs a custom EfficientDet-based architecture and is capable of finding comets in single images in near real time, rather than requiring multiple epochs as with traditional methods. The system achieves state-of-the-art performance with 99% recall, a 0.01% false-positive rate, and a 1-2 pixel rms error in the predicted position. We report the initial results of the Tails efficiency evaluation in a production setting on the data of the ZTF Twilight survey, including the first AI-assisted discovery of a comet (C/2020 T2) and the recovery of a comet (P/2016 J3 = P/2021 A3).
AB - We present Tails, an open-source deep-learning framework for the identification and localization of comets in the image data of the Zwicky Transient Facility (ZTF), a robotic optical time-domain survey currently in operation at the Palomar Observatory in California, USA. Tails employs a custom EfficientDet-based architecture and is capable of finding comets in single images in near real time, rather than requiring multiple epochs as with traditional methods. The system achieves state-of-the-art performance with 99% recall, a 0.01% false-positive rate, and a 1-2 pixel rms error in the predicted position. We report the initial results of the Tails efficiency evaluation in a production setting on the data of the ZTF Twilight survey, including the first AI-assisted discovery of a comet (C/2020 T2) and the recovery of a comet (P/2016 J3 = P/2021 A3).
UR - https://www.scopus.com/pages/publications/85108180910
UR - https://www.scopus.com/pages/publications/85108180910#tab=citedBy
U2 - 10.3847/1538-3881/abea7b
DO - 10.3847/1538-3881/abea7b
M3 - Article
AN - SCOPUS:85108180910
SN - 0004-6256
VL - 161
JO - Astronomical Journal
JF - Astronomical Journal
IS - 5
M1 - 218
ER -