Abstract: The generation of high-quality images from textual descriptions is a challenging task in computer vision and natural language processing. The goal of text-to-image synthesis, a current topic ...