Multi-Density Sketch-to-Image Translation Network

Sketch-to-image (S2I) translation plays an important role in image synthesis and manipulation tasks, such as photo editing colorization. Some specific S2I translations, including sketch-to-photo sketch-to-painting, can be used powerful tools the art design industry. However, previous methods only support with a single level of density, which gives less flexibility to users for controlling input sketches. In this work, we propose first multi-level density sketch-to-image framework, allows sketch cover wide range from rough object outlines microstructures. Moreover, tackle problem noncontinuous representation sketches, project into continuous latent space, then linearly controlled by parameter. This conveniently control densities sketches generation images. our method has been successfully verified on various datasets different applications, face editing, multi-modal translation, anime colorization, providing coarse-to-fine levels controls these applications.