ml4ai / ggan

Texture generation using a Graph Generative Adversarial Network and Differentiable Rendering

Source code for the paper:

Texture Generation Using a Graph Generative Adversarial Network and Differentiable Rendering

https://link.springer.com/chapter/10.1007/978-3-031-25825-1_28

https://arxiv.org/pdf/2206.08547.pdf

Download ShapeNet car dataset from: https://shapenet.org/

requirements:

python==3.8
pytorch==1.9.0+cu111
pytorch3d==0.6.0

train [ggan model]:

sh Experiments/gnn_kraken.sh

test [ggan model]:

sh Experiments/test.sh

Note: The source code contains multiple files used to train other models

Note: The optimization problem is complicated [changing hyperparameters slightly may result in a huge change in the generated texture quality]

If you use the source code please cite the following paper:

@inproceedings{dharma2023texture,
  title={Texture Generation Using a Graph Generative Adversarial Network and Differentiable Rendering},
  author={Dharma, KC and Morrison, Clayton T and Walls, Bradley},
  booktitle={Image and Vision Computing: 37th International Conference, IVCNZ 2022, Auckland, New Zealand, November 24--25, 2022, Revised Selected Papers},
  pages={388--401},
  year={2023},
  organization={Springer}
}