/PWCC

Pixel-wise Color Constancy via Smoothing Techniques in Multi-Illuminant Scenes

Primary LanguagePythonMIT LicenseMIT

PWCC: Pixel-wise Color Constancy via Smoothing Techniques in Multi-Illuminant Scenes

This repository includes the implementation details of Pixel-wise Color Constancy (PWCC) proposed in arXiV:

Pixel-wise Color Constancy via Smoothing Techniques in Multi-Illuminant Scenes

SCC_Pipeline drawio

PWCC learns a pixel-wise mapping, referred to as an illuminant map, between input and ground truth images. This mapping provides a white-balancing operation to the input image so that color casts can be minimized.

Conda environment setup:

# In Linux Environment:
$ while read requirement; do conda install --yes $requirement || pip install $requirement; done < envcc.txt

Citation

If you use the provided method in this repository, please cite the following paper:


@misc{entok2024pixelwise,
      title={Pixel-Wise Color Constancy via Smoothness Techniques in Multi-Illuminant Scenes}, 
      author={Umut Cem Entok and Firas Laakom and Farhad Pakdaman and Moncef Gabbouj},
      year={2024},
      eprint={2402.02922},
      archivePrefix={arXiv},
      primaryClass={cs.CV}
}

Downloading the Dataset

This paper uses LISMI Dataset. Please follow the instructions of the study employed this dataset from here: LSMI Dataset Download