Abstract
We present a model for generating postage stamp images of synthetic Fanaroff-Riley Class I and Class II radio galaxies suitable for use in simulations of future radio surveys such as those being developed for the Square Kilometre Array. This model uses a fully-connected neural network to implement structured variational inference through a variational auto-encoder and decoder architecture. In order to optimise the dimensionality of the latent space for the auto-encoder we introduce the radio morphology inception score (RAMIS), a quantitative method for assessing the quality of generated images, and discuss in detail how data pre-processing choices can affect the value of this measure. We examine the 2-dimensional latent space of the VAEs and discuss how this can be used to control the generation of synthetic populations, whilst also cautioning how it may lead to biases when used for data augmentation.
| Original language | English |
|---|---|
| Pages (from-to) | 3351-3370 |
| Number of pages | 20 |
| Journal | Monthly Notices of the Royal Astronomical Society |
| Volume | 503 |
| Issue number | 3 |
| Early online date | 15 Mar 2021 |
| DOIs | |
| Publication status | Published - May 2021 |
Funding
DJB gratefully acknowledges support from Science and Technology Facilities Council (STFC) and the Newton Fund through the Development in Africa through Radio Astronomy (DARA) Big Data program under grant ST/R001898/1. AMS gratefully acknowledges support from an Alan Turing Institute AI Fellowship EP/V030302/1. MB gratefully acknowledges support from the University of Manchester STFC Centre for Doctoral Training (CDT) in Data Intensive Science, grant number ST/P006795/1. FP gratefully acknowledges support from STFC and IBM through the iCASE studentship ST/P006795/1
Keywords
- astro-ph.IM
- astro-ph.GA
- methods: statistical
- surveys
- radio continuum: galaxies