The problem of source separation in two dimensions is studied in this paper. The problem is formulated in the Bayesian framework. The sources are modelled as MRFs to accommodate for the spatially correlated structure of the sources, which we exploit for separation in 2D. The difficulty of working analytically with general Gibbs distributions is overcome by using an approximate density. In this work, the Gibbs distribution is modelled by the product of directional Gaussians. The sources are estimated by Maximuma- Posteriori estimation using the approximate density as the prior. At each iteration of the MAP estimation, an annealing schedule is used for approximate density. This annealing schedule aids the algorithm to converge the global extremum. The mixing matrix is found by Maximum Likelihood estimation.
Source separation in images via MRFs with variational approximation
Kuruoglu E E;
2007
Abstract
The problem of source separation in two dimensions is studied in this paper. The problem is formulated in the Bayesian framework. The sources are modelled as MRFs to accommodate for the spatially correlated structure of the sources, which we exploit for separation in 2D. The difficulty of working analytically with general Gibbs distributions is overcome by using an approximate density. In this work, the Gibbs distribution is modelled by the product of directional Gaussians. The sources are estimated by Maximuma- Posteriori estimation using the approximate density as the prior. At each iteration of the MAP estimation, an annealing schedule is used for approximate density. This annealing schedule aids the algorithm to converge the global extremum. The mixing matrix is found by Maximum Likelihood estimation.File | Dimensione | Formato | |
---|---|---|---|
prod_91695-doc_131634.pdf
solo utenti autorizzati
Descrizione: Source separation in images via MRFs with variational approximation
Tipologia:
Versione Editoriale (PDF)
Dimensione
525.4 kB
Formato
Adobe PDF
|
525.4 kB | Adobe PDF | Visualizza/Apri Richiedi una copia |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.