$μ$NCA: Texture Generation with Ultra-Compact Neural Cellular Automata
This enables highly efficient texture generation for graphics applications, though it is incremental in model compression.
The paper tackles the problem of example-based procedural texture synthesis by training ultra-compact Neural Cellular Automata (NCA) models, achieving models with as few as 68 parameters that can generate complex textures comparable to hand-engineered programs.
We study the problem of example-based procedural texture synthesis using highly compact models. Given a sample image, we use differentiable programming to train a generative process, parameterised by a recurrent Neural Cellular Automata (NCA) rule. Contrary to the common belief that neural networks should be significantly over-parameterised, we demonstrate that our model architecture and training procedure allows for representing complex texture patterns using just a few hundred learned parameters, making their expressivity comparable to hand-engineered procedural texture generating programs. The smallest models from the proposed $μ$NCA family scale down to 68 parameters. When using quantisation to one byte per parameter, proposed models can be shrunk to a size range between 588 and 68 bytes. Implementation of a texture generator that uses these parameters to produce images is possible with just a few lines of GLSL or C code.