INR-st

Controllable Style Transfer via Test-time Training of Implicit Neural Representation

This is the implementation of the paper "Controllable Style Transfer via Test-time Training of Implicit Neural Representatione" by Sunwoo Kim, Youngjo Min, Younghun Jung and Seungryong Kim.

For more information, check out the paper and project page on [arXiv, project page].

Overall Architecture

Our model "Controllable Style Transfer via Test-time Training of Implicit Neural Representation" is illustrated below:

Environment Settings

git clone https://github.com/KU-CVLAB/INR-st.git 
cd INR-st

pip install -r requirements.txt

Optimization

bash ./run_train.sh

Inference

Default

bash ./run_size_interpolation.sh

Super Resolution

bash ./run_size_control.sh

Gradation

bash ./run_gradation.sh

Region-wise stylization

bash ./run_mask.sh

BibTeX

If you find this research useful, please consider citing:

@article{kim2022inrst,
  title = {Controllable Style Transfer via Test-time Training of Implicit Neural Representation},
  author = {Kim, sunwoo and Youngjo, Min and Younghun, Jung and Kim, Seungryong},
  journal = {arXiv preprint arXiv:2210.07762},
  year = {2022},
}