DAGAN: A Domain-Aware Method for Image-to-Image Translations

Figure 3

Background model overview; this model aims to learn a cross-domain transformation between both background inputs.