WebtoonMe: A Data-Centric Approach for Full-Body Portrait Stylization

Jihye Back*
Seungkwon Kim*
Namhyuk Ahn    (* denotes equal contribution)
NAVER WEBTOON AI
[Paper]
[Demo - Coming Soon]



Abstract

Full-body portrait stylization, which aims to translate portrait photography into a cartoon style, has drawn attention recently. However, most methods have focused only on converting face regions, restraining the feasibility of use in real-world applications. A recently proposed two-stage method expands the rendering area to full bodies, but the outputs are less plausible and fail to achieve quality robustness of non-face regions. Furthermore, they cannot reflect diverse skin tones. In this study, we propose a data-centric so- lution to build a production-level full-body portrait stylization system. Based on the two-stage scheme, we construct a novel and advanced dataset prepa- ration paradigm that can effectively resolve the aforementioned problems. Experiments reveal that with our pipeline, high-quality portrait stylization can be achieved without additional losses or architectural changes.

Video


Method

An overview of the data preparation stream.


Results

Ablation Study.


Paper


Jihye Back, Seungkwon Kim, Namhyuk Ahn.
WebtoonMe: A Data-Centric Approach for Full-Body Portrait Stylization
In SIGGRAPH Asia 2022, Technical Communications. [Paper]
[Bibtex]


Citation

If you find this useful for your research, please use the following.
@inproceedings{back2022webtoonme,
  title={WebtoonMe: A Data-Centric Approach for Full-Body Portrait Stylization},
  author={Back, Jihye and Kim, Seungkwon and Ahn, Namhyuk},
  journal={arXiv preprint arXiv:2210.10335},
  year={2022}
}