Patch-based dictionary learning for sparse image approximationMS2

We propose a new regularization method for sparse representation and denoising of seismic data. Our approach is based on a sparse data representation in a learned dictionary and a similarity measure for image patches that is evaluated using the Laplacian matrix of a graph. We propose dictionary learning algorithms based on clustering and singular value decomposition. We also consider a similarity measure for the local geometric structures of the data.

This presentation is part of Minisymposium “MS2 - Interpolation and Approximation Methods in Imaging (4 parts)
organized by: Alessandra De Rossi (University of Torino) , Costanza Conti (University of Firenze) , Francesco Dell'Accio (University of Calabria) .

Gerlind Plonka (University of Goettingen)
image compression, image reconstruction, image representation, inverse problems, nonlinear optimization