Project page | Video | Paper
Official implementation of DCT-Net for Portrait Stylization.
DCT-Net: Domain-Calibrated Translation for Portrait Stylization,
Yifang Men1, Yuan Yao1, Miaomiao Cui1, Zhouhui Lian2, Xuansong Xie1,
1DAMO Academy, Alibaba Group, Beijing, China
2Wangxuan Institute of Computer Technology, Peking University, China
In: SIGGRAPH 2022 (TOG) arXiv preprint
(2022-07-07) The paper is available now at arxiv(https://arxiv.org/abs/2207.02426).
(2022-08-08) cartoon function can be directly call from pythonSDK of modelscope.
(2022-08-08) The pertained model and infer code of 'anime' style is available now. More styles coming soon.
- python 3
- tensorflow (>=1.14)
- easydict
- numpy
- both CPU/GPU are supported
git clone https://github.com/menyifang/DCT-Net.git
cd DCT-Net
A quick use with python SDK
- Installation:
conda create -n dctnet python=3.8
conda activate dctnet
pip install tensorflow
pip install "modelscope[cv]" -f https://modelscope.oss-cn-beijing.aliyuncs.com/releases/repo.html- Downloads:
python download.py- Inference:
python run_sdk.pypython run.pyFace detector and aligner are adapted from Peppa_Pig_Face_Engine and InsightFace.
If you find this code useful for your research, please use the following BibTeX entry.
@inproceedings{men2022dct,
title={DCT-Net: Domain-Calibrated Translation for Portrait Stylization},
author={Men, Yifang and Yao, Yuan and Cui, Miaomiao and Lian, Zhouhui and Xie, Xuansong},
journal={ACM Transactions on Graphics (TOG)},
volume={41},
number={4},
pages={1--9},
year={2022},
publisher={ACM New York, NY, USA}
}