Scale Mixtures of Gaussians and the Statistics of Natural Images
Abstract
The statistics of photographic images, when represented using multiscale (wavelet) bases, exhibit two striking types of non(cid:173) Gaussian behavior. First, the marginal densities of the coefficients have extended heavy tails. Second, the joint densities exhibit vari(cid:173) ance dependencies not captured by second-order models. We ex(cid:173) amine properties of the class of Gaussian scale mixtures, and show that these densities can accurately characterize both the marginal and joint distributions of natural image wavelet coefficients. This class of model suggests a Markov structure, in which wavelet coeffi(cid:173) cients are linked by hidden scaling variables corresponding to local image structure. We derive an estimator for these hidden variables, and show that a nonlinear "normalization" procedure can be used to Gaussianize the coefficients.
Cite
Text
Wainwright and Simoncelli. "Scale Mixtures of Gaussians and the Statistics of Natural Images." Neural Information Processing Systems, 1999.Markdown
[Wainwright and Simoncelli. "Scale Mixtures of Gaussians and the Statistics of Natural Images." Neural Information Processing Systems, 1999.](https://mlanthology.org/neurips/1999/wainwright1999neurips-scale/)BibTeX
@inproceedings{wainwright1999neurips-scale,
title = {{Scale Mixtures of Gaussians and the Statistics of Natural Images}},
author = {Wainwright, Martin J. and Simoncelli, Eero P.},
booktitle = {Neural Information Processing Systems},
year = {1999},
pages = {855-861},
url = {https://mlanthology.org/neurips/1999/wainwright1999neurips-scale/}
}