Flow matching technique generates realistic polycrystalline microstructures
$C_4$-Equivariant Flow Matching on Anisotropic Power-Diagram Graphs for Microstructure Generation
Machine Learning
Summary
Getting real microstructure data, which shows how tiny crystals are arranged in metals, usually needs expensive and special machines. The authors present a way to create realistic images of these microstructures using a type of artificial intelligence that treats the structures like graphs with directional properties. Their method understands rotations, so turning the input noise rotates the generated pattern too. They can also guide the AI without retraining it to produce specific metal types, like welds or 3D-printed alloys.
What this means in practice
- •For materials engineers: create realistic synthetic microstructures to test and design new metal alloys without costly physical experiments
- •For computer graphics developers: generate high-resolution microstructure textures with guaranteed rotational consistency for scientific visualization and simulation
Authors
Dawid Lipinski, Jixiang Qing, Henry Moss
Abstract
Acquiring realistic microstructure data through Electron Backscatter Diffraction (EBSD) is costly and time consuming, often relying on specialised equipment. As microstructures strongly influence material properties, generating realistic samples is essential for modelling the behaviour of polycrystalline materials. We introduce a generative model for synthesising realistic polycrystalline microstructures using flow matching and graph neural networks. By representing microstructures as anisotropic power diagrams, our model learns a compact geometric parametrisation and can render generated samples at arbitrary pixel resolution. A $C_4$-equivariant architecture incorporates rotational symmetry directly into the model, ensuring that rotations of the input noise produce corresponding rotations of the generated microstructure. We also demonstrate how training-free guidance can be used to generate complex microstructures, based on user defined objective function. In particular, we generate microstructures resembling a copper weld, cast metal slab, 3D-printed stainless steel and heterogeneous lamella titanium.