Codage arithmetique pour la description d'une distribution

TL;DR

This paper introduces an efficient model selection tool called the RIC information criterion, based on predictive adaptive arithmetic coding and MDL, and extends it to non-parametric distribution estimation with practical image histogram applications.

Contribution

It develops the RIC criterion for model selection using arithmetic coding and MDL, and extends coding techniques to non-parametric distribution estimation.

Findings

RIC criterion effectively selects models in experiments.

Extension to non-parametric estimation applied to image histograms.

Demonstrates practical utility in image analysis.

Abstract

Using predictive adaptive arithmetic coding and the Minimum Description Length principle, we derive an efficient tool for model selection problems : the RIC information criterion. We then present an extension of these coding techniques to non-parametrical estimation of a distribution and illustrate it on the gray scales histogram of an image. Key-words : Information criteria, MDL, model selection, non-parametrical estimation, histograms.