Authors
Philipp Henzler, Niloy J. Mitra, Tobias Ritschel
University College London; Adobe Research
Portals
Abstract
We propose a generative model of 2D and 3D natural textures with diversity, visual fidelity and at high computational efficiency. This is enabled by a family of methods that extend ideas from classic stochastic procedural texturing (Perlin noise) to learned, deep, non-linearities. The key idea is a hard-coded, tunable and differentiable step that feeds multiple transformed random 2D or 3D fields into an MLP that can be sampled over infinite domains. Our model encodes all exemplars from a diverse set of textures without a need to be re-trained for each exemplar. Applications include texture interpolation, and learning 3D textures from 2D exemplars.
Overview
Our approach has two steps. The first embeds the exemplar into a latent space using an encoder. The second provides sampling at any position by reading noise fields at that position and combining them using a learned mapping to match the exemplar statistics.