Paper: https://arxiv.org/abs/2005.07702

This project takes on the problem of transferring the style of cartoon images to real-life photographic images by implementing previous work done by CartoonGAN. We trained a Generative Adversial Network(GAN) on over 60 000 images from works by Hayao Miyazaki at Studio Ghibli.

1. Install Anaconda from https://www.anaconda.com/

2. Install pytorch at: https://pytorch.org/get-started/locally/

3. Install dependencies:

   python -m pip install tqdm pillow numpy matplotlib opencv-python
   

4. For predicting videos you will also need ffmpeg

All training code can be found in experiment.ipynb

Predict by running predict.py.

Example:

python predict.py -i C:/folder/input_image.png -o ./output_folder/output_image.png

Predictions can be made on images, videos or a folder of images/videos.

Image #	Original	CartoonGAN	GANILLA	Our implementation
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20