 |
Communication dans un congrès
SEAFLOOR CLASSIFICATION USING STATISTICAL MODELING OF WAVELET SUBBANDS
Abstract : This paper deals with the classification of textured seafloor images recorded by
sidescan sonar. To address this problem, a supervised classification approach based on
the Bayesian framework is proposed. In this way, the textured images are characterized
through parametric probabilistic models of the wavelet coefficients. The generalized
Gaussian distribution (GGD), which is a well-established model to characterize the
marginal distributions of the wavelet subbands, is considered. However, to take into
account the joint statistics of wavelet coefficients, we also consider the Gaussian copula
based multivariate generalized Gaussian model (GC-MGG). A supervised learning
context is adopted for the classification stage by using a probabilistic k-Nearest
Neighbors classifier. Each textured image will be represented by its GGD or GC-MGG
estimated parameters and given a collection of training images the Kullback-Leibler
divergence is used to estimate the similarity between a test image and seafloor classes.
Experiments on real sonar textured images are proposed to highlight the interest of this
approach
https://hal.archives-ouvertes.fr/hal-01171474
Contributeur : Annick Billon-Coat <>
Soumis le : vendredi 3 juillet 2015 - 17:11:23
Dernière modification le : mercredi 24 juin 2020 - 16:19:24
Contributeur : Annick Billon-Coat <>
Soumis le : vendredi 3 juillet 2015 - 17:11:23
Dernière modification le : mercredi 24 juin 2020 - 16:19:24