Bayesian Color Constancy with Non-Gaussian Models
Abstract
We present a Bayesian approach to color constancy which utilizes a non- Gaussian probabilistic model of the image formation process. The pa- rameters of this model are estimated directly from an uncalibrated image set and a small number of additional algorithmic parameters are chosen using cross validation. The algorithm is empirically shown to exhibit RMS error lower than other color constancy algorithms based on the Lambertian surface reflectance model when estimating the illuminants of a set of test images. This is demonstrated via a direct performance comparison utilizing a publicly available set of real world test images and code base.
Cite
Text
Rosenberg et al. "Bayesian Color Constancy with Non-Gaussian Models." Neural Information Processing Systems, 2003.Markdown
[Rosenberg et al. "Bayesian Color Constancy with Non-Gaussian Models." Neural Information Processing Systems, 2003.](https://mlanthology.org/neurips/2003/rosenberg2003neurips-bayesian/)BibTeX
@inproceedings{rosenberg2003neurips-bayesian,
title = {{Bayesian Color Constancy with Non-Gaussian Models}},
author = {Rosenberg, Charles and Ladsariya, Alok and Minka, Tom},
booktitle = {Neural Information Processing Systems},
year = {2003},
pages = {1595-1602},
url = {https://mlanthology.org/neurips/2003/rosenberg2003neurips-bayesian/}
}