Skip to content

Latest commit

 

History

History
24 lines (12 loc) · 966 Bytes

README.md

File metadata and controls

24 lines (12 loc) · 966 Bytes

SliceGAN

To use SliceGAN open run_slicegan and edit information as requested. SliceGAN requires a single 2D training image of an isotropic microstructure, or three 2D images taken at perpendicular angles of an anisotropic microstructure. Images can be colour, grayscale or n-phase.

Use 'python run_slicegan 1' to train a new generator or 'python run_slicegan 0' to generate and save an example .tif file

Development

If you are interested in trying new architectures, see the networks.py file

To adjust the training parameters or algorithm, see train.py

To add a new preprocessing method e.g for a different training datatype, see preproccessing.py.

Results

Versions

v1.1 release: DOI

v1.0 release: DOI