On Finding Gray Pixels

Matlab Code accompanying the paper On Finding Gray Pixels. [arXiv]

From left to right: (a) input image. (b) computed grayness index GI. darker blue indicates higher degree of grayness. (c) the N % most gray pixels rendered using the corresponding pixel color (greenish) in (a). (d) estimated illumination color. (e) ground truth color. (f) corrected image using (d).

If you use grayness index code or results, please consider citing following paper:

@inproceedings{qian2019cvpr,
  title={On Finding Gray Pixels},
  author={Qian, Yanlin and K{\"a}m{\"a}r{\"a}inen, Joni-Kristian and Nikkanen, Jarno and Matas, Jiri},
  booktitle={IEEE International Conference of Computer Vision and Pattern Recognition},
  year={2019}
}

If you use classic gray pixel code or results, please consider citing following paper:

@inproceedings{yang2015efficient,
  title={Efficient illuminant estimation for color constancy using grey pixels},
  author={Yang, Kai-Fu and Gao, Shao-Bing and Li, Yong-Jie},
  booktitle={CVPR},
  year={2015}
}

If you use mean-shifted gray pixel code or results ([here] ), please consider citing following paper:

@inproceedings{qian2019vissap,
  title={Revisiting Gray Pixel for Statistical Illumination Estimation},
  author={Qian, Yanlin and Purtuz, Said and Nikkanen, Jarno and K{\"a}m{\"a}r{\"a}inen, Joni-Kristian and Matas, Jiri},
  booktitle={International Conference of Computer Vision Theory and Applications},
  year={2019}
}