Wide-field submillimetre surveys have driven many major advances in galaxy evolution in the past decade, but without extensive follow-up observations the coarse angular resolution of these surveys limits the science exploitation. This has driven the development of various analytical deconvolution methods. In the last half a decade Generative Adversarial Networks have been used to attempt deconvolutions on optical data. We have developed an auto-encoder with a novel loss function to overcome this problem in the submillimeter wavelength range. This approach works on Herschel SPIRE 500 μm COSMOS data, with the superresolving target being the JCMT SCUBA-2 450 μm observations of the same field. The auto-encoder reproduces the JCMT SCUBA-2 images with high fidelity. This is quantified through the point source fluxes and positions, the completeness, and the purity.
Slides: in PDF
Video: https://youtu.be/uHlycbhOkGM
| Subject : | : | Oral presentation |
| Comment | : | This presentation is based on the work done for the MNRAS article "Superresolving Herschel imaging: a proof of concept using Deep Neural Networks" (https://doi.org/10.1093/mnras/stab2195). |
| Topics | : | Cosmology |
| Topics | : | Imaging |
| Topics | : | Concepts and algorithms |
| Topics | : | Contributed |
| Keywords | : | Neural Networks ; Herschel Space Observatory ; JCMT ; Sub ; mm galaxies |
| PDF version | : |  PDF version |
