DeepSource: Point Source Detection using Deep Learning

Point source detection at low signal-to-noise is challenging for astronomical surveys, particularly in radio interferometry images where the noise is correlated. Machine learning is a promising solution, allowing the development of algorithms tailored to specific telescope arrays and science cases. We present DeepSource - a deep learning solution - that uses convolutional neural networks to achieve these goals. DeepSource enhances the Signal-to-Noise Ratio (SNR) of the original map and then uses dynamic blob detection to detect sources. Trained and tested on two sets of 500 simulated 1 deg x 1 deg MeerKAT images with a total of 300,000 sources, DeepSource is essentially perfect in both purity and completeness down to SNR = 4 and outperforms PyBDSF in all metrics. For uniformly-weighted images it achieves a Purity x Completeness (PC) score at SNR = 3 of 0.73, compared to 0.31 for the best PyBDSF model. For natural-weighting we find a smaller improvement of ~40% in the PC score at SNR = 3. If instead we ask where either of the purity or completeness first drop to 90%, we find that DeepSource reaches this value at SNR = 3.6 compared to the 4.3 of PyBDSF (natural-weighting). A key advantage of DeepSource is that it can learn to optimally trade off purity and completeness for any science case under consideration. Our results show that deep learning is a promising approach to point source detection in astronomical images.

[1]  G. Bruce Berriman,et al.  Astrophysics Source Code Library , 2012, ArXiv.

[2]  Yoshua Bengio,et al.  Convolutional networks for images, speech, and time series , 1998 .

[3]  Jian Sun,et al.  Deep Residual Learning for Image Recognition , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[4]  M. Kunz,et al.  Bayesian inference for radio observations , 2015, 1501.05304.

[5]  O. Smirnov,et al.  The MeqTrees software system and its use for third-generation calibration of radio interferometers , 2010, 1101.1745.

[6]  D. J. Saikia,et al.  A 325-MHz GMRT survey of the Herschel-ATLAS/GAMA fields , 2013, 1307.4590.

[7]  India,et al.  An Australia Telescope survey for CMB anisotropies , 2000 .

[8]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.

[9]  Edward J. Kim,et al.  Star-galaxy Classification Using Deep Convolutional Neural Networks , 2016, ArXiv.

[10]  O. Lahav,et al.  PHOTOMETRIC SUPERNOVA CLASSIFICATION WITH MACHINE LEARNING , 2016, 1603.00882.

[11]  T. Murphy,et al.  wsclean: an implementation of a fast, generic wide-field imager for radio astronomy , 2014, 1407.1943.

[12]  Jürgen Schmidhuber,et al.  Deep learning in neural networks: An overview , 2014, Neural Networks.

[13]  Joana Frontera-Pons,et al.  Unsupervised feature-learning for galaxy SEDs with denoising autoencoders , 2017, 1705.05620.

[14]  A. Shulevski,et al.  The ASKAP/EMU Source Finding Data Challenge , 2015, Publications of the Astronomical Society of Australia.

[15]  Ben Hoyle,et al.  Measuring photometric redshifts using galaxy images and Deep Neural Networks , 2015, Astron. Comput..

[16]  Cheng Soon Ong,et al.  Radio Galaxy Zoo:Claran– a deep learning classifier for radio morphologies , 2018, Monthly Notices of the Royal Astronomical Society.

[17]  Barnabás Póczos,et al.  Enabling Dark Energy Science with Deep Generative Models of Galaxy Images , 2016, AAAI.

[18]  Melvin M. Varughese,et al.  Nonparametric Transient Classification using Adaptive Wavelets , 2015, 1504.00015.

[19]  E. Rol,et al.  PySE: Software for extracting sources from radio images , 2018, Astron. Comput..