In this paper, we address the task of layout-to-image translation, which aims to translate an input semantic layout to a realistic image. One open challenge widely observed in existing methods is the ...