Awesome

High-definition Cloth Retouching (CRHD-3K) Dataset

High-definition Cloth Retouching (CRHD-3K) is a high-quality dataset for cloth retouching, which aims to remove the wrinkles, creases, and other blemishes on the clothes to make them look more smooth and beautiful. It includes 3,022 pairs of raw and retouching photos, where the raw photos were collected from Unsplash and the retouching process was accomplished by a team of professional image editors. The dataset is released as part of the following paper:

ABPN: Adaptive Blend Pyramid Network for Real-Time Local Retouching of Ultra High-Resolution Photo<br> Biwen Lei, Xiefan Guo, Hongyu Yang, Miaomiao Cui, Xuansong Xie, Di Huang<br> In CVPR 2022<br>

DAMO Academy, Alibaba Group, Hangzhou, China

Paper | Supp

We will release the code of ABPN soon, before that, if you are interested in skin retouching, you may refer to modelscope, in which we released the models of portrait skin retouching for application. In one of the models, we leveraged the concept of blend layers in ABPN and achieved good results.

Samples

Overview

All data (CRHD-3K) can be downloaded below.

Agreement

• All files in the CRHD-3K dataset are available for non-commercial research purposes only.
• You agree not to reproduce, duplicate, copy, sell, trade, resell or exploit for any commercial purposes, any portion of the images and any portion of derived data.

Citation

If you use this dataset for your research, please cite our paper.

@inproceedings{lei2022ABPN,
  title={ABPN: Adaptive Blend Pyramid Network for Real-Time Local Retouching of Ultra High-Resolution Photo},
  author={Biwen, Lei and Xiefan, Guo and Hongyu, Yang and Miaomiao, Cui and Xuansong,  Xie and Di, Huang},
  booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition},
  year={2022}
}

Contact

Should you have any questions, please contact me via biwen.lbw@alibaba-inc.com.