Controlling Material Appearance by Examples

Despite the ubiquitous use of materials maps in modern rendering pipelines, their editing and control remains a challenge. In this paper, we present an example-based material method to augment input based on user-provided photos. We train tileable version MaterialGAN leverage its prior guide appearance transfer, optimizing latent space using differentiable rendering. Our transfers micro meso-structure textures user provided target(s) photographs, while preserving structure quality material. show our methods can existing maps, increasing realism or generating new, visually appealing materials.