Afficher la couverture 'Exact Rényi and Kullback-Leibler Divergences Between Multivariate t-Distributions | Bouhlel, Nizar' en grand format
/5

0 avis

Exact Rényi and Kullback-Leibler Divergences Between Multivariate t-Distributions

Archive ouverte

Bouhlel, Nizar | Rousseau, David

Edité par CCSD ; Institute of Electrical and Electronics Engineers -

International audience. In this letter, we propose a closed-form expression of the Rényi divergence (RD) of order β between two zero-mean real multivariate t-distributions (MTDs). Such distribution has been deployed in several signal and image processing applications where heavy-tailed distribution is well-suited. Based on the computation of the multiple integral involved in the RD, the expression of the divergence is provided without resorting to the conventional time-consuming Monte Carlo (MC) integration technique. In addition, the Kullback-Leibler divergence (KLD) is deduced from RD. Finally, a comparison is made between the MC method and the numerical value of the RD expression to show how the former gives close approximations to the latter.

Suggestions

Du même auteur

Multivariate Statistical Modeling for Multitemporal SAR Change Detection Using Wavelet Transforms and Integrating Subband Dependencies

Archive ouverte | Bouhlel, Nizar | CCSD

International audience. In this article, we propose a new method for automatic change detection in multitemporal fully polarimetric synthetic aperture radar (PolSAR) images based on multivariate statistical wavelet ...

Parameter Estimation of the Normal Ratio Distribution with Variational Inference

Archive ouverte | Bouhlel, Nizar | CCSD

International audience. In this paper, we propose a new method for parameter estimation of the probability density function of the photosystem II (PSII) index in chlorophyll fluorescence imaging. The PSII index is m...

Online Bayesian Adaptive Sampling for Nonlinear Model: Application to Plant Phenotyping

Archive ouverte | Mercier, Félix | CCSD

International audience. Online adaptive temporal sampling for the monitoring of a nonlinear process of real interest in plant phenomics is developed. We propose to quantify the uncertainty at each sample to be acqui...

Chargement des enrichissements...