Awesome
White-box facial image cartoonizaiton
- Pytorch implementation of a simple facial image cartoonizaiton netowrk
- This work is an improvement based on CVPR2020 White-box cartoonizaiton paper
- project page | paper | twitter | zhihu | bilibili
Some results
<img src="results/00050.jpg" width="512px"/>
<img src="results/00051.jpg" width="512px"/>
<img src="results/00055.jpg" width="512px"/>
<img src="results/00064.jpg" width="512px"/>
<img src="results/00066.jpg" width="512px"/>
<img src="results/00067.jpg" width="512px"/>
<img src="results/00069.jpg" width="512px"/>
<img src="results/00089.jpg" width="512px"/>
<img src="results/00091.jpg" width="512px"/>
<img src="results/00094.jpg" width="512px"/>
<img src="results/00096.jpg" width="512px"/>
<img src="results/00097.jpg" width="512px"/>
Installation
- Install pytorch, you don't even need cuda and gpu
- Opencv, numpy and tqdm are needed, I think you already have them
Inference with Pre-trained Model
- Crop your facial image to 256*256 pixel
- Save them in the 'images' folder
- Run 'inference.py', output will be saved in 'results' folder
Training code
- Sorry, it's not available now, I am still working on improving this work
- If I am lucky enough, maybe this method will appear in a paper
License