Spatial Generative Adversarial Networks: Tensorflow

Very Brief Model Description

SGANs can generate sample textures of arbitrary size that look strikingly similar - but not exactly the same - compared to a single (or several) source image(s).

• SGANs can be thought of as a convolutional roll-out of Radford et al.'s deep convolutional generative adversarial networks for texture synthesis
• the fully convolutional nature allows for real-time generation of high resolution images
• the method can fuse multiple source images and is highly scalable w.r.t. to output texture size and allows for generation of tiled textures

Training the Model

All model parameters can be viewed by

python run.py --help

We have tried to be consistent with the original implementation https://github.com/ubergmann/spatial_gan, however some alterations have been made. To run the training, one needs to either adjust the parameters in run.py and use

python run.py

or

python run.py --data_dir /path_to_data_dir/

In order to use metric regularization described https://arxiv.org/pdf/1612.02136.pdf run

python run.py --data_dir /path_to_data_dir/ --reg True

Generating results

For regular SGAN

python run.py --is_train False model_dir /path/to/dir/with/checkpoints

for metric regularized SGAN

python run.py --is_train False model_dir /path/to/dir/with/checkpoints --reg True

Results

This model was trained on a google maps image of barcelona, and yields a texture image like e.g. this

and for metric regularized

License

The MIT License (MIT)

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.