Showing 1–1 of 1 results for author: Balbás, D

Machine learning approach to the detection of point sources in maps of the CMB temperature anisotropies

Authors: P. Diego-Palazuelos, R. B. Barreiro, P. Vielva, D. Balbás, M. López-Caniego, D. Herranz, B. Casaponsa

Abstract: We propose a machine learning approach to the blind detection of extragalactic point sources on maps of the temperature anisotropies of the cosmic microwave background. Using realistic simulations of the microwave sky as seen by Planck, we train a convolutional neural network (CNN) that solves source detection as an image segmentation problem. We divide the sky into regions of progressively increa… ▽ More We propose a machine learning approach to the blind detection of extragalactic point sources on maps of the temperature anisotropies of the cosmic microwave background. Using realistic simulations of the microwave sky as seen by Planck, we train a convolutional neural network (CNN) that solves source detection as an image segmentation problem. We divide the sky into regions of progressively increasing Galactic foreground intensity and independently train specialized CNNs for each region. This strategy leads to promising levels of completeness and reliability, with our CNN substantially outperforming traditional detection methods like the matched filter in regions close to the Galactic plane. △ Less

Submitted 28 February, 2023; originally announced February 2023.

Comments: 3 pages, 1 figure. Proceedings of the "Machine Learning: a giant leap towards space discovery in the era of peta and exabyte scale surveys" Symposia of the 2022 Annual meeting of the European Astronomical Society. To be published in Memorie della SAIt