Applying machine learning for transient detection/classification in difference images.
Product Goal: a general-purpose difference image analysis pipeline that can identify multiply-imaged transients using machine learning.
Inputs: difference images as FITS files. Associated source catalogs from the static-sky (template) images and the difference images.
Outputs: categorization score for any transient candidates in the image using three classes 0 - not a real transient 1 - single transient 2 - multiply-imaged transient candidate
Target datasets: Strong lensing systems observed with Las Cumbres Observatory and the Hubble Space Telescope; simulated images for the Rubin Observatory LSST and Roman Space Telescope.
Makes use of Gaia data: https://www.cosmos.esa.int/web/gaia-users/credits