Visually prompted keyword localisation Poster presented at the Deep Learning Indaba 2022 by Aletta S E Nortje