Neural Color Operators for Sequential Image Retouching
Tsinghua University
Baidu Inc.
Tsinghua University
Baidu Inc.
Abstract
We propose a novel image retouching method by modeling the retouching process as performing a sequence of newly introduced trainable \emph{neural color operators}. The neural color operator mimics the behavior of traditional color operators and learns pixelwise color transformation while its strength is controlled by a scalar. To reflect the homomorphism property of color operators, we employ equivariant mapping and adopt an encoder-decoder structure which maps the non-linear color transformation to a much simpler transformation (i.e., translation) in a high dimensional space. The scalar strength of each neural color operator is predicted using CNN based strength predictors by analyzing global image statistics. Overall, our method is rather lightweight and offers flexible controls. Experiments and user studies on public datasets show that our method consistently achieves the best results compared with SOTA methods in both quantitative measures and visual qualities.
Paper
BibTeX
Please cite our work if you use code or data from this site.
@inproceedings{wang2022neurop,
author = {Wang, Yili and Li, Xin and Xu, Kun and He, Dongliang and Zhang, Qi and Li, Fu and Ding, Errui},
title = {Neural Color Operators for Sequential Image Retouching},
year = {2022},
isbn = {978-3-031-19800-7},
publisher = {Springer-Cham},
url = {https://doi.org/10.1007/978-3-031-19800-7_3},
doi = {10.1007/978-3-031-19800-7_3},
booktitle = {Computer Vision – ECCV 2022: 17th European Conference, Tel Aviv, Israel, October 23–27, 2022, Proceedings, Part XIX},
numpages = {14},
}Results
MIT-Adobe FiveK-Lite
MIT-Adobe FiveK-Dark
