Abstract: According to various embodiments, an artificial intelligence framework capable of neural synthesis of tileable textures is provided. Using a non-tileable texture input, such as single or multi-layer texture stacks, high-quality tileable textures can be generated that match the appearance of the input samples, while increasing their spatial resolution. Embodiments are provided that leverage a latent space approach in a generative network for synthesizing seamlessly tileable textures, which can maintain semantic consistency in boundaries between tiles. A tileabilty metric is provided as feedback to improve and optimize the tileability of the outout texture, for example using a sampling algorithm that can generate high-resolution, artifact-free tileable textures. In embodiments, a convolutional discriminator is provided for detecting artifacts in the synthesized textures by locally estimating the quality of the synthesized maps.

Abstract: According to various embodiments, a framework for transferring micro-scale properties of any kind (geometry, optical properties, material parameters resolution, etc.) of a material, to larger images of the same material is provided. The framework includes a machine learning model capable of learning to perform this transfer from a set of micro-scale images of one or more patches of the material and micro-scale properties of the material. The machine learning framework transfers the micro-scale properties to images of the same material, regardless of its size, resolution, illumination or geometry.